Sixth International Symposium on Cavitation

CAV2006, Wageningen, The Netherlands, September 2006

ANALYSES OF UNSTEADY CAVITATION IN ROCKET FEED SYSTEMS AND
FLOW CONTROL ELEMENTS

Vineet Ahuja and Ashvin Hosangadi
Combustion Research and Flow Technology, Inc.
Pipersville, PA USA
vineet@craft-tech.com

ABSTRACT
Feed systems and flow control elements in
rocket propulsion systems and testing facilities are
constantly subject to dynamic events that lead to
unsteady fluctuations in pressure and mass flow. Such
events can also be accompanied by cavitation,
resonance, system vibration leading to premature system
shutdown and/or catastrophic failure. In this paper, we
examine the multi-phase instability mechanisms in feed
system components through the use of high-fidelity
dynamic computational simulations. Our numerical
framework incorporates a cryogenic cavitation model
that accounts for thermodynamic effects related to
cavitation processes such as leading edge temperature
depression and evaporative cooling. Additionally, we
account for unsteady effects through the use of a bubble
dynamics model that tracks the growth and collapse of
vapor clouds. We discuss two problems related to feed
system components, a pump discharge orifice and a
coupled turning duct venturi system, that represent very
diverse cavitation regimes. Detailed analyses of the
physical phenomena in the two systems are provided
along with fundamental excitation frequencies.

INTRODUCTION
Experimental testing of rocket engines and
associated test articles are carried out in facilities that
have complex flow paths that comprise of a network of
valves, pipes, arterial branches, orifices and junctions.
The complexity in the flow path often leads to flow
induced instabilities caused by a variety of factors
including vortex shedding, cavitation, turbulence and
flow separation. In such complex flow environments it is
not uncommon for the system to experience severe
pressure fluctuations from the coupling between the test
article/engine and the valve and feed systems that
comprise the test facility leading to premature shutdown
of the tests. Furthermore, such deleterious pressure
fluctuations are not limited to rocket engine test facilities
and have been observed in feed systems onboard the
space shuttle and the Delta 4 launch systems. Cavitation
related pressure fluctuations were thought to be largely
responsible for pressure fluctuations in the oxidizer
feedline of the Delta 4 system leading to spurious sensor
readings and the premature shutdown of the RS-68
engine. Large scale pressure fluctuations have also
contributed to structural damage such as the cracks
observed in the space shuttle liquid hydrogen feed liners.
Computational simulations can play an integral role
in supporting testing and developmental activities by
identifying and characterizing these instabilities.
However, the diversity of flow regimes and instability
mechanisms place very stringent requirements on any
computational framework that could be used for such
analyses. For example, the identification of dominant
frequencies associated with flow instabilities in such
systems requires high order numerics, advanced
turbulence modeling capabilities, sophisticated grid
topologies to resolve local physics in complex
geometries, embedded models for unsteady cavitation,
capture thermal effects in cryogenic fluids, and dynamic
motion of feed system components such as valves. The
instability mechanisms in rocket testing facilities can be
classified into three distinct classes: (a)
hydrodynamic/fluid dynamic instabilities [1] that are
predominantly attributed to fundamental flow physical
mechanisms such as vortex shedding, turbulence, etc.
largely associated with complex structural configurations
such as valve housing, plug shapes, manifolds and bends
in the piping system. (b) flow transients associated with
valve timing and valve scheduling [2]. Such transients
can play an important role especially during
startup\shutdown and valve response can be critical for
safe and reliable operation. (c) multi-phase instabilities
such as cavitation related instability mechanisms that
become especially important with cryogenic working
fluids due to reduced liquid to vapor density ratios and
strong evaporative cooling effects resulting in local
temperature fluctuations that can couple with the
primary cavitation instability.
In the past, we have tailored our multi-element
unstructured CFD solver to carry out simulations of
instability mechanisms that lead to valve chatter and
operational stall in complex valve and feed system
components [3]. Computational analyses have been
performed for a variety of structurally complex valve
systems such as the pressure regulator valve, split body
valve and gaseous control valves. We have also
extended our multi-element unstructured framework to
include timing studies for transient valve operation by
including valve movement and capturing the flow
instabilities associated with it. In this paper, we discuss
Paper No. 80, CAV2006 Ahuja and Hosangadi 1

simulations of cavitating instabilities in feed system
components and control elements used in the E-1 test
facility at NASA Stennis Space Center (SSC).
Cavitation related instabilities are endemic in liquid
rocket feed systems and test facilities. Since these flow
regimes primarily consist of cryogenic fluids that operate
in close proximity to the critical temperature, there are
substantial thermal effects and property variations
associated with such flows. Our cavitation models [4]
account for the coupling of thermodynamic processes
with the cavitation processes and have accurately
captured such effects as leading edge pressure and
temperature depression due to evaporative cooling, and
frothy cavitation zones that are the hallmark of cavitated
regions in cryogenic fluids. Furthermore, with the
incorporation of an unsteady cavitation model [5] we are
able to predict amplitudes and frequencies of dynamic
pressure loads and track bubble clouds sheared off
cavities due to the interaction of reentrant jets in the
cavity closure region.
The two different feed system components
considered here are: (a) orifice and (b) flow control
venturi that is attached to a 90 degree turning duct.
These two components represent the two disparate
regimes of cavitating flows: the orifice is representative
of traveling cavitation class of flows commonly
observed in tip vortices of propeller blades, whereas the
venturi represents a sheet cavitation type problem with
periodic shedding of vaporous clouds from this well
developed cavity. The simulations are performed with a
well validated cryogenic cavitation model [4] that takes
into account evaporative cooling and other associated
thermal effects.
In the next section, we briefly discuss our multi-
phase formulation, followed by a section on our
unsteady cavitation model. This is followed by a
discussion of our simulations of the cavitating
instabilities related to the orifice and the flow control
venturi used in test facilities at NASA SSC. We finish
the paper with a brief summary.

MULTI-PHASE EQUATION SYSTEM
We give a very brief overview of the basic
multiphase equation system here and refer the reader to
Hosangadi and Ahuja [4] for more details. The equation
system is written solved in a pressure based form as:

v
=S+D
G
z

+
0
0
0
0
t
t fg
k
S
m
m h
S
S

=








( / )
v
Q Q =
v
Q E F
t x y


+ + (1)

(2)
2

m
m
m
m
v
g g g
m m m
m
m
u
P
u u P
v uv
w uw
Q E
u
h h u
k ku
u

+








= =













The vectors Qv, E and S are given above. The
matrix defines the transformation from the
conservative to primitive variables and may further be
preconditioned to obtain an efficient time-marching
scheme.
The source term for the vapor phase arises due to
cavitation where m
t
is the net rate of vapor mass
generation (or condensation), and the corresponding
source term for the energy equation is given as m
t
h
fg

where h
fg
is the change in enthalpy resulting from the
phase change and is a function of the local fluid
temperature. These phase change source terms are
discussed in a later section.
The mixture density, enthalpy, and vapor porosity are
related by the following relations locally in a given cell
volume:
= +
m g g L L
(3)


= +
m m g g g L L L
h h h
1
g L
(4)
= +
g
h
(5)
where
g
,
L
are the physical material densities,
while and
L
h are the sensible enthalpy of the vapor
and liquid phase respectively, and in general are
functions of both the local temperature and pressure. In
our study here, these properties were generated from the
Standard thermodynamic database 12 available from
NIST for pure fluids [6]. The thermodynamic properties
of the fluid where specified using the saturation values
from the table corresponding to the local temperature of
the fluid. Equations (1)-(2) represent a stiff system with
large variations in the acoustic speed that are a function
of the local multi-phase composition. Preconditioning
techniques are used to overcome this stiffness and obtain
an efficient numerical scheme [4].

UNSTEADY CAVITATION MODEL USING CLOUD
SURFACE AREA EQUATION
The new, unsteady cavitation model developed
incorporates formulations for bubble dynamics (e.g.
Rayleigh-Plesset equation) as an integral part of dense
cloud cavitation. By drawing on our experience in dense
spray combustion [7], we track the surface area
associated with the cavitation cloud as it evolves
spatially and temporally. We note that this Eulerian
procedure has been formulated within the premise of a
dense bubble cloud where a large number of bubbles are
present. Bubbles comprising the cloud will be
characterized by their Sauter mean radius, which
preserves the ratio of the surface area and volume for the
bubbly cloud. Thus, the source term for cavitation will
have two independent factors controlling it: a) the net
surface area given by the bubble Sauter mean radius, and
b) the rate of change of the radius, which may be
specified using the Rayleigh-Plesset equation.
The additional equation (shown in 1-D form for
discussion purposes) for the bubble surface area is given
as:
Paper No. 80, CAV2006 Ahuja and Hosangadi 2


g g
t
S S u
s
t x

+ =

(6)
Here s
t
is the source terms to the cloud surface area
equation derived in a manner consistent with the
Rayleigh Plesset equation and is defined later. For
clarity, we repeat the vapor mass conservation, which
was already included as part of the original multi-phase
system (Eqn. (2)),

g g g g
t
u
m
t x

+ =

g g

(7)
The dependent variables S and may be written
as
(8)
2
*4.0* * , *4.0
g g
S N r N = =
3
* * /3.0 r
3* /
g g
r S
where N is the number density of bubbles, and r the
Sauter mean radius of bubbles at each grid point. Thus,
the Sauter mean radius, r, that conserves cloud surface
area and volume is defined as, =
2.0* /
t g g t
t g t
m S r
. Note that
no restriction is placed on the vapor volume fraction
being dilute. Furthermore we do not restrict the number
density of bubbles to be a constant in each cell, and
allow them to convect as governed by the
hydrodynamics.
The source terms for the net vapor mass transfer and
surface area change are given by

s S r r

r
t
t
r
=
=
(9)
where is the rate of change of the radius. To specify
the rate of change of radius , the Rayleigh-Plesset
Equation could be specified as follows:

2
2
2
3 1
( )
2


+ =

i
l
d r dr
r P
dt r dt
2 4

dr
P
r dt
(10)
However this would require the solution of an
additional transport equation for the solution of (dr/dt).
In our study a simplified and approximate form for the
radius change is used:

1/ 2
( )

v
l
P P
P
v
P
v
P
1/ 2
2
3

=


abs dr
dt
(11)
where the sign of the radius change term depends on
whether the bubble is growing or decaying and is
dictated by the sign of the term. Here, is the
vapor pressure, while
CAVITATION ANALYSES OF FLOW THROUGH
ORIFICE
An orifice is routinely used in testing facilities to
step down pressure as well as for flow control and in this
section we analyze the cavitating instability that sets up
in the testing facility (NASA SSC) due to a pump
discharge orifice. The flow rate of liquid hydrogen
through the orifice is 130 lbs/sec at an operating
temperature at 21.7 K. The inlet pipe has a diameter of 6
inches and the orifice throat diameter is 3.26 inches
(with an inlet radius of 0.75 inches). A back pressure
corresponding to 65 psia is maintained on the outlet end
of the configuration and the corresponding vapor
pressure of liquid hydrogen at the operating temperature
is 21.755 psia. Since the flow through the orifice pertains
to a cavitating instability, the simulations in this case
were carried out with unsteady bubbly cavitation model
that was discussed earlier. This cavitation model as
discussed previously solves an additional equation for
the net surface area of the vapor cloud and replicates the
bubble dynamics as represented by the Rayleigh-Plesset
equation. Incorporating this cavitation model in the
simulations permits tracking of vapor clouds as they
convect through the piping downstream of the orifice.
The flow was seeded with micron sized vapor bubbles
and an initial void fraction of 10
-6
.
A snapshot of the instantaneous axial velocity
distribution is shown in Figure 1(a), which indicates the
formation of a primary jet as flow accelerates to
negotiate the orifice. It should be noted that this jet is
representative of a very high Reynolds Number flow
since cryogens such as liquid hydrogen typically have
very low viscosity. Figure 1(b) shows the vorticity
associated with the fringes of this jet and fluctuations in
vorticity lead to pressure oscillations. Figure 1(c)
depicts a snapshot of the pressure distribution with
spatially periodic patterns of low pressure in the pipe
downstream of the orifice. Vorticity production at the lip
of the orifice leads to unsteady shedding and the periodic
formation of these pockets of low pressure. The
sequence of pressure distribution in Figure 2 shows the
periodic formation of these low frequency patterns
downstream of the orifice. When the pressure in these
pockets falls below the vapor pressure cavitation sets in
leading to the formation of vapor clouds that grow and
convect downstream (Figure 3). Interestingly, low
temperature is found in the traveling vapor clouds
(Figure 4) indicating that they are prone to condensation
events and pressure excursions as a result of phase
change.
Surprisingly, large vapor clouds do not shed off a
dominant sheet-like cavity from the lip of the orifice as
seen in conventional cavitation regimes such as ogives
and hydrofoils. Rather, the interaction of low pressure
associated with vorticity in the wake of the orifice with a
bubbly vapor trail leads to the growth and formation of
large vapor clouds almost a diameter aft of the orifice
lip. This phenomenon is not unlike the growth and
formation of vapor clouds in the wake region or tip
vortex of a propeller blade.
P is the instantaneous pressure at
the grid node point at which the cavitation source term is
being computed, and
l
is the liquid density. We again
emphasize that Eqn. (11) provides the phase change for
an individual bubble only while the net mass transfer is
given by Eqn. (9) by accounting for the total surface area
of bubbles present at a given spatial grid location.
Paper No. 80, CAV2006 Ahuja and Hosangadi 3

(a)
(b)
(c)
Figure 1. Instantaneous snapshots of flow solution
through the orifice (a) axial velocity (b) vorticity (c)
pressure distribution.


(a)
(b)
(c)
(d)
(e)
(f)
(g)
(h)

Figure 2. Sequence of pressure distributions showing
periodic formation of pockets of low pressure.
(a)
(b)
(c)
(d)
(e)
(f)
(g)
(h)

Figure 3. Sequence of void fraction plots showing the
convecting of vapor clouds.

The physics behind tip vortex cavitation was poorly
understood due to the lack of a vapor cavity on the
propeller blade surface. The mechanics of traveling
cavitation and the interaction of vorticity in the wake
with a bubbly vapor trail to form large vapor clouds, as
noticed here can provide important insights into
Paper No. 80, CAV2006 Ahuja and Hosangadi 4

observed cavitation phenomena in ship and submarine
propeller systems.

Figure 4. Instantaneous snapshot of temperature
distribution

The shedding of the vapor clouds in our case is a
fairly periodic phenomenon as evidenced by the
instantaneous void fraction distribution seen in Figure 3.
Furthermore, these clouds are formed in regions of high
vorticity in the shear layer associated with the primary
jet. The growth and development of the vapor clouds be
better explained by the sequence of instantaneous void
fraction contour plots in a blown up region around the
orifice (Figure 5). Here, we see a small cavity formed at
the lip of the orifice with a well-defined gas-liquid
bubbly wake. As the wake encounters a region of low
pressure downstream of the orifice, it leads to sudden
expansion and growth of the vapor cloud. The formation
of the low pressure region a pipe diameter downstream
of the orifice corresponds to the sudden expansion and
growth in the vapor cloud. Figure 6 shows the Fourier
decomposition of pressure traces of two numerical
probes located 5 inches and 10 inches downstream of the
orifice along the piping wall. Both probes show large
scale pressure fluctuations attributed to the highly
dynamic processes of formation and collapse of the
vapor clouds. The frequency spectrum of the pressure
oscillations indicates a cavitation shedding frequency of
80 Hz is excited in both probes with significant energy in
a higher overtone of 320 Hz.





Figure 5. Blowup of region close to orifice showing sequence of instantaneous void fraction distributions.
80 Hz
Hz
7 Hz
325
154
80 Hz
320 Hz
800 Hz

(a) 5 inches downstream of orifice (b) 10 inches downstream of orifice
Figure 6. Fourier decomposition of numerical pressure traces.
Paper No. 80, CAV2006 Ahuja and Hosangadi 5

SIMULATIONS OF TURNING DUCT VENTURI
SYSTEM
The venturi is primarily used as a flow control
device in various propulsive applications and test
facilities. Control in the venturi is exerted through a
cavitation mechanism that under certain operating
conditions can make the mass flow rates through the
venturi insensitive to pressure fluctuations. Previously,
[3] axisymmetric simulations of a cavitating venturi with
liquid oxygen as the operating fluid were carried out
with cryogenic cavitation finite rate mode [4] in
CRUNCH CFD

and variable thermodynamic properties

extracted from a NIST databank. This particular venturi
has a throat diameter of 2.94 inches and an inlet diameter
of 10 inches with a sharply contoured contraction region
and a long divergent section. In this case, the calculated
discharge coefficient (0.96) compared favorably with the
experimental discharge coefficient (0.95). Moreover,
these simulations were able to explain the physical
mechanisms behind cavitation based control that is
exerted by the venturi. The simulations demonstrated
that cavitation inception initiates in the throat region
(Figure 7a) where the local pressure falls below the
saturation vapor pressure corresponding to the bulk
temperature. As the simulation evolves, the
thermodynamic effects of cavitation become more
pronounced leading to evaporative cooling near the
cavity region (Figure 7b). This results in a temperature
depression, and consequently a depression in the
saturation pressure at the leading edge of the cavity.
These results are consistent with the findings of research
carried out by Stahl and Stepanoff [8], Ruggeri and
Moore [9], Hord [10] and Brennen [11],[12].
Furthermore, our simulations reveal a vaporous mixture
in the entire venturi cross-section just downstream of the
throat. The vapor fraction through most of the cavity is
less than 0.5 indicating a highly diffused cavity Figure
7a). Sarosdy and Acosta [13], have made a similar
observations of investigating of frothy, diffusive cavities
consisting of many bubbles with Freon 113. Moreover,
the mass flow through the venturi gets choked due to the
presence of the vapor cloud near the throat of the
venturi. Any variation in back pressure is not
propagated to the inlet due to a substantial reduction in
speed of sound through the vaporous cavity.
The simulations carried out here relate to the same
cavitating venturi (discussed above) in a LOX feedline
that is connected to a 90 degree turning duct. The
coupled turning duct and venturi simulations were
performed primarily to look at dynamic effects in the
system and understand the associated physics. For
computational tractability a 2-D configuration of the
system was assumed as a axisymmetric 3-D assumption
is not possible due to the presence of the turning duct.
The freestream conditions for this case consisted of a
freestream velocity of 33.06 m/s and a density of
1132.12 kg/m
3
at a temperature of 92 K.
The vapor pressure at the freestream temperature
corresponded to 17.69 psi. A sequence of instantaneous
void fraction distributions are shown in Figure 8 that
depict the process of breakup of the sheet cavity that
forms along the walls of the venturi and consequently
the convection of the vapor clouds. It should be noted
that the asymmetry of the cavitated zones along the
upper and lower walls of the cavity is a direct
consequence of the coupling between the venturi and the
turning duct. A strong reentrant jet develops in the
closure region of the cavity that shears large vapor
clouds and results in the oscillation of the cavity length.
(a)
(b)
Figure 7. Void Fraction and Temperature
Distribution for the LOX Venturi.

The sequence of instantaneous snapshot of the
pressure distribution (Figure 9) in the system reveals a
highly dynamic pressure field both in the turning duct
(near the bend) and in the venturi (in the cavity closure
region). First, flow through the bend is highly unstable
leading to pressure oscillations that travel through the
length of the turning duct to the entrance of the venturi.
Second, pressure waves are also seen to travel
downstream of the venturi in the discharge duct as a
result of the shedding and condensation of vapor clouds
in the venturi. Third, pockets of high pressure are seen in
the aft end of the cavity/venturi that represents
condensation fronts of convected vapor clouds.
Pronounced thermodynamic effects that are
characteristic of cavitating regimes in cryogenic fluids
can be seen in the sequence of instantaneous temperature
distribution snapshots (Figure 10).
Evaporative cooling effects that result in a
temperature depression at the leading edge of the cavity
can be clearly seen here (Figure 10). Furthermore, we
find that the temperature in the core of the vapor clouds
that are shed off the main sheet cavity is also
significantly lower than the freestream temperature.
Such strong temperature (and consequently, vapor
pressure) depressions have the potential for creating
condensation fronts leading to large pressure excursions
and a highly dynamic flowfield. The spectral content of
two numerical probes (one downstream of the bend,
other downstream of sheet cavity region) show very
different excitation frequencies.


Paper No. 80, CAV2006 Ahuja and Hosangadi 6






Figure 8. Instantaneous snapshot of vapor void fraction showing shedding of vapor clouds.





Figure 9. Instantaneous pressure distribution in the turning duct venturi system.






Figure 10. Instantaneous temperature distribution in the turning duct venturi system.
Paper No. 80, CAV2006 Ahuja and Hosangadi 7

The probe in the turning duct shows a low
frequency dominant mode (100 Hz) related to
unsteadiness associated with boundary layer separation
along the inner wall just aft of the 90 degree bend with
pressure oscillations of about 10%. In contrast, the
probe in the venturi shows significant energy in higher
frequencies (476 Hz primary excitation mode) and large
scale pressure fluctuations (Figure 11).

100 Hz
550 Hz

(a) Numerical probe in turning duct downstream of bend
476 Hz
1973 Hz
2722 Hz

(b) Numerical probe in venturi aft of the sheet cavity
Figure 11. Frequency Spectra for numerical probes
in the turning duct venturi system.

SUMMARY
The effort in this paper revolved around the
application of our unsteady CFD framework with
embedded sub-models for turbulence and cavitation to
predict dynamic pressure loads and their associated
frequencies in feed system components. Our cavitation
model has been developed around fundamental bubble
dynamics (represented by the Rayleigh-Plesset equation)
and tracks the growth and collapse of bubble clouds by
solving additional transport equations for the
conservation of net surface area and void fraction of the
bubble cloud. Furthermore, our equation system is cast
in a form that permits accurate representation of acoustic
wave propagation in cryogenic and two-phase regimes
with phase change and temperature fluctuations strongly
coupled to pressure and flow variations.
The primary aim of this paper is to understand the
key physical processes that lead to cavitation instabilities
in two different feed system components by carrying
unsteady cavitating simulations. These components are
regularly used in test facilities and include a step down
orifice (E-1 Cell 2) and a coupled turning duct venturi
(E-1 Cell 1) system. Both simulations were performed
for cryogenic fluids (liquid hydrogen and liquid oxygen)
in flow regimes where thermodynamic effects are
significant at operating conditions used in the facility.
Simulations of the orifice show a periodic shedding of
vapor clouds from the lip of the orifice and the formation
of these clouds is directly linked to vorticity production
from the orifice. The growth and formation of these
vapor clouds is unlike traditional shedding of vapor
clouds from sheet cavities and is accompanied by large
scale pressure and temperature fluctuations.
Furthermore, fundamental frequencies associated with
the cavitation instability and the shedding process was
captured as part of the simulation process on numerical
probes placed on the wall downstream of the orifice at
different locations. Fourier decomposition of the
pressure traces indicates substantial energy in both a
fundamental mode of 80Hz and a 320Hz overtone.
Simulations were also performed for a coupled
turning duct venturi system operating in liquid oxygen.
Here, the dynamic process of formation, shedding and
convection of large scale vapor clouds in the venturi was
captured as part of the simulation. Unlike the instability
mechanism in the orifice, a sheet like cavity evolved
along the walls of the venturi near the throat. A strong
reentrant jet formed in the closure region of the cavity
and was responsible for breakup of the cavity and the
shedding of vapor clouds aft of the cavity. As a
consequence of the interaction of the reentrant jet with
the cavity, the cavity length oscillates due to periodic
shearing off from the action of the reentrant jet.
Furthermore, these simulations showed a strong coupling
between thermodynamic effects and the cavitation
process with temperature depression at the leading edge
of the cavity and in the core of vapor clouds shed due to
the interaction of the reentrant jet and the cavity. Low
temperature in the clouds results in a lowering of the
local vapor pressure leading to condensation events in
the migrating vapor clouds and strong fluctuations in
pressure.
Coupling the turning duct to the venturi in our
simulations has important consequences: Firstly, there is
asymmetry observed in the cavities that form near the
throat of the venturi along the upper and lower walls.
This is primarily from the flow asymmetry feeding in
from the turning duct into the venturi. Secondly, under
choked conditions the venturi is unresponsive to small
pressure oscillations in the turning duct and the
cavitation instability in the venturi is the dominant
instability mechanism. Numerical probes along the
walls in the turning duct venturi system show small
pressure fluctuations in the turning duct (10% of mean
pressure) attributed to vortex shedding processes in the
bend region, while much larger pressure oscillations
(100% of mean pressure) are seen in the venturi.
Frequency spectra of probes in the turning duct and the
venturi indicate very different modes are excited in the
disparate parts of the system. Vortex shedding from the
bend in the turning duct is responsible for small
amplitude low frequency excitation whereas a high
frequency excitation causing large scale fluctuations is
seen in the venturi primarily due to the cavitation
shedding and condensation process.

ACKNOWLEDGMENTS
We acknowledge funding for this work through a
NASA SBIR, Contract Number NNS04AA08C funded
Paper No. 80, CAV2006 Ahuja and Hosangadi 8

by NASA Stennis Space Center. Dr. Peter Sulyma was
the technical monitor and his input is gratefully
acknowledged.
REFERENCES
[1] Ahuja, V., Hosangadi, A., Shipman, J ., and Cavallo, P.A..,
Simulations of Instabilities in Complex Valve and Feed
Systems, AIAA-2006-4758, 42
nd
AIAA/ASME/SAE/
ASEE J oint Propulsion Conference & Exhibit, 9 - 12 J ul
2006, Sacramento, CA.
[2] Cavallo, P.A., Hosangadi, A., and Ahuja, V., Transient
Simulations of Valve Motion in Cryogenic Systems,
AIAA Paper No. 2005-5152, 35th Fluid Dynamics
Conference and Exhibit, Toronto, Canada, J une 2005.
[3] Ahuja, V., Hosangadi, A., Shipman, J ., Daines, R., and
Woods, J ., Multi-Element Unstructured Analyses of
Complex Valve Systems, Accepted for Publication,
Journal of Fluids Engineering, May 2006.
[4] Hosangadi, A. and Ahuja, V., Numerical Study Of
Cavitation In Cryogenic Fluids, Journal of Fluids
Engineering, Vol. 127, pp. 267-281, March 2005.
[5] Hosangadi, A., and Ahuja, V., A New Unsteady
Model for Dense, Cloud Cavitation in Cryogenic
Fluids, Paper No. AIAA-2005-5347, 17
th

Computational Fluids Dynamic Conference,
Toronto, Ontario, CA, J un. 6-9, 2005.
[6] Lemmon, E.W., Peskin, A.P., McLinden, M.O., and
Friend, D.G., (2000) NIST Thermodynamic and
Transport Properties of Pure Fluids, NIST Standard
Reference Database 12, Users Guide.
[7] Hosangadi, A., Sinha, N., and Dash, S.M., "A Unified
Hyperbolic Interface Capturing Scheme for Gas/Liquid
Flows," AIAA-97-2081, 13th AIAA CFD Conferences,
Snowmass, CO, J une 29-J uly 2, 1997.
[8] Stahl, H.A., and Stepanoff, A.J ., Thermodynamic
Aspects of Cavitation in Centrifugal Pumps, ASME J.
Basic Eng., Vol. 78, pp. 1691-1693, 1956.
[9] Ruggeri, S.R., and Moore, R.D., Method for Prediction
of Pump Cavitation Performance for Various Liquids,
Liquid Temperature, and Rotation Speeds, NASA TND-
5292, 1969.
[10] Hord, J ., Cavitation In Liquid Cryogens, NASA CR-
2156, J anuary 1973.
[11] Brennen, C.E., The Dynamic Behavior and Compliance
of a Stream of Cavitating Bubbles, Journal of Fluids
Engineering, pp. 533-542, Vol. 95, 1973.
[12] Brennen, C.E., Hydrodynamics of Pumps, Oxford
University Press, 1994.
[13] Sarosdy L.R. and Acosta A.J ., Note on Observations of
Cavitation in Different Fluids, Paper No. 60-WA-83,
ASME Winter Annual Meeting, New York, Nov. 27- Dec.
2, 1960.
Paper No. 80, CAV2006 Ahuja and Hosangadi 9