Wei Review Cavitation PoF 2022

REVIEW ARTICLE | OCTOBER 21 2022
Cavitation in cryogenic fluids: A critical research review 

Special Collection: Cavitation
Aibo Wei (魏爱博) ; Lianyan Yu (余潋滟); Limin Qiu (邱利民); Xiaobin Zhang (张小斌) 
Physics of Fluids 34, 101303 (2022)

https://doi.org/10.1063/5.0102876

View Export
Online Citation
Articles You May Be Interested In
Cavitation flow characteristics on the surface of hydrofoil with microjet structure

Physics of Fluids (July 2024)
Characterization of cavitation zone in cavitating venturi flows: Challenges and road ahead
Physics of Fluids (November 2023)
Data-driven modal decomposition of transient cavitating flow

Physics of Fluids (November 2021)
22 November 2024 04:35:04

Physics of Fluids REVIEW scitation.org/journal/phf
Cavitation in cryogenic fluids: A critical research

review
Cite as: Phys. Fluids 34, 101303 (2022); doi: 10.1063/5.0102876
Submitted: 13 June 2022 . Accepted: 31 July 2022 .
Published Online: 21 October 2022
Aibo Wei (魏爱博), Lianyan Yu (余潋滟), Limin Qiu (邱利民), and Xiaobin Zhang (张小斌)a)
AFFILIATIONS
Institute of Refrigeration and Cryogenic, Zhejiang University, Hangzhou 310027, China
Note: This paper is part of the special topic, Cavitation.

a)
Author to whom correspondence should be addressed: [email protected]
ABSTRACT
Cavitation occurs as the fluid pressure is lower than the vapor pressure at a local thermodynamic state and may result in huge damage to the
hydraulic machinery. Cavitation in cryogenic liquids is widely present in contemporary science, and the characteristics of cryogenic
cavitation are quite different from those of water due to thermal effects and strong variations in fluid properties. The present paper reviews
recent progress made toward performing experimental measurements and developing modeling strategies to thoroughly investigate
22 November 2024 04:35:04

cryogenic cavitation. The thermodynamic properties of cryogenic fluids are first analyzed, and different scaling laws for thermal effects
estimation are then introduced. As far as cryogenic cavitation experimental research is concerned, the progress made in the cavitation
visualization and cavity dynamics and the synchronous measurements of the multi-physical field are mainly introduced. As for the study on
numerical simulation of cryogenic cavitation, the commonly used cavitation models and turbulence models are, respectively, classified and
presented, and the modifications and improvements of the cavitation model and turbulence model for thermal effect modeling of cryogenic
cavitation are examined. Then, several advances of critical issues in cryogenic fluid cavitation research are reviewed, including the influences
of thermal effects, unsteady shedding mechanisms, cavitation–vortex interactions, and cavitation-induced vibration/noise. This review offers
a clear vision of the state-of-the-art from both experimental and numerical modeling viewpoints, highlights the critical study developments
and identifies the research gaps in the literature, and gives an outlook for further research on cryogenic cavitation.
Published under an exclusive license by AIP Publishing. https://doi.org/10.1063/5.0102876
NOMENCLATURE Pr Prandtl number

Q heat flux (W/m2)
a thermal diffusivity (m2/s) R mass transfer rate [kg/(m3 s)] bubble radius (m)
B B-factor Re Reynolds number
c speed of sound (m/s) S entropy (J K1)
Cp pressure coefficient SE energy source term
D length scale (m) t time (s)
f frequency (Hz); mass fraction T shedding cycle (s); temperature (K)
F fluid force (N) u velocity (m/s)
h convective heat transfer coefficient [W/(m2 K)] sensible U velocity scale
enthalpy (J) V vapor volume (m3)
k thermal conductivity [W/(m K)] x, y, z coordinate axes
L length (m) specific heat capacity [J/(kg K)]
Lev latent heat of vaporization (J/kg)
m mass (kg) Greek symbols
Nb bubble number density
Nu Nusselt number a volume fraction
p pressure (Pa) c surface tension (N/m)
Phys. Fluids 34, 101303 (2022); doi: 10.1063/5.0102876 34, 101303-1

Published under an exclusive license by AIP Publishing
d boundary layer/cavity thickness (m) everything in cryogenic engineering revolves around cryogenic fluids.
R thermodynamic parameter (m/s3/2) Common cryogenic liquids of concern include liquid nitrogen (LN2),
l dynamic viscosity (Pa s) liquid hydrogen (LH2), liquid oxygen (LOX), liquid methane (LCH4)/
q density (kg/m3) liquid natural gas (LNG), and liquid helium (LHe), which are often
r cavitation number used in numerous large industries with different needs such as aero-
s subgrid-scale stress space, tool manufacturing, air separation, chemical engineering, cryo-
x vorticity genic wind tunnels, and laboratory work. Very diverse fluid
machinery, meanwhile, is unavoidably employed in cryogenic engi-
Superscripts/subscripts neering such as pumps, inducers, turbines, nozzles, valves, etc., in
c cavity which the phenomenon of cryogenic cavitation rifely consists as a
i, j, k direction of Cartesian coordinates result of the wide-ranging pressure variations along with the flow.1–5
l liquid Proverbially, cryogenic cavitation acts as one of the most important
m mixture issues of concern in the industry because of the considerable impact,
t turbulent which can result in component failure,6 reduced efficiency,7 vibration
u, v, w velocity components instability of operation,8 and so on. Various typical engineering appli-
v vapor cation scenarios including the use of cryogenic fluids and the relevant
1 reference value fluid machinery where cryogenic cavitation often occurs are illustrated
þ condensation in Fig. 1.
- evaporation One of the most important and most widely exploited cryogenic
liquids is the liquid rocket propellant, which serves as the popular fuel
Abbreviations for heavy-lifting space launch vehicles worldwide.9,10 Typically, a
high-power-density cryogenic turbopump with the smallest possible
DCM dynamic cavitation model
size is applied to supply the propellants to the combustion chamber
LES large eddy simulation
operated at fairly high pressure, which demands relatively high impel-
LH2 liquid hydrogen
ler speeds due to the size constraint on the turbopump. A negative
LN2 liquid nitrogen
22 November 2024 04:35:04

static pressure region is likely to form due to the high speeds and inevi-
LOX liquid oxygen
tably induce the cavitation of the propellant, which then results in
NPSH net positive suction head
VN2 vaporous nitrogen severe performance deterioration of the turbopump. To improve sys-
tem performance, the inducer is attached to the turbopump to pressur-
ize the cryogenic propellant sufficiently to guarantee the normal
I. INTRODUCTION operation of the turbopump and, therefore, increase its efficiency.
Cryogenics is finding important applications over an extraordi- However, the inducer also suffers from cavitation, especially at blade
narily diverse range of engineering and technology, and almost surfaces, and the cavitation-induced flow instabilities can affect the
FIG. 1. Typical engineering applications of cryogenic fluids.
Phys. Fluids 34, 101303 (2022); doi: 10.1063/5.0102876 34, 101303-2

rotor dynamics of the impeller or even result in launch failures. For LOX (90 K) are 27.6, 11.6, and 10.4 kPa/K, respectively, and
example, the failure of the launch of H-II Rocket was mainly attributed only 0.12 kPa/K for water (298 K).
to the resonance of the liquid hydrogen pump inducer with one of the (2) The liquid–vapor density ratio of the cryogenic liquid (qv/ql) is
inducer blades aroused by the vibration stress due to swirling cavita- more than 100 times smaller than the same ratio for water,
tion, and the unexpected vibration caused fatigue fracture of the which means more liquids are needed to vaporize to sustain a
inducer blade, which then triggered the failure of the liquid hydrogen cavity, and a greater latent heat is required to be supplied by the
turbopump. A similar cause associated with the cavitation-induced liquid to the interface during the vaporization process. The
instability was also well ascertained for the failure of Ariane 5. ensuing need for increased heat transfer decreases the tempera-
Consequently, great efforts have been made to obtain a thorough ture of the local fluid.
understanding of the cavitating flow characteristics inside the inducer (3) The thermal conductivity of cryogenic fluid is much smaller.
in order to improve the turbopump’s cavitation performance. Hence, there will be a larger temperature gradient inside the
In addition to the critical effect of cryogenic cavitation on the cryogenic fluid, that is, there will be a larger temperature drop
operation of the liquid propulsion system, the widely influential effects near the cavitation zone under the same heat flux condition,
of cavitation in cryogenic fluids also exist in other many large, complex which will then result in a large reduction in the local saturation
industrial systems, ranging from the important function in lowering vapor pressure.
the measurement accuracy of a flowmeter to the magnifying effect for Compared with water at room temperature, the physical proper-
vibrational instability of transportation pipe system. It is, therefore, ties (temperature-dependent) of cryogenic fluids are more thermo-
crucial to establish profound and comprehensive knowledge of the sensitive, and the substantial reductions in the cavity temperature are
characteristics and mechanism of cryogenic cavitation, and the signifi- especially present due to the latent heat of vaporization, which is taken
cance of an integrated study based on experimental inquiry and to the liquid around the cavitation area whose temperature is then
computational modeling to deal with this critical and difficult issue is lowered. In the case of water cavitation at room temperature, the varia-
self-explanatory. tion of temperature caused by the evaporation phase is near 0.01 K,
while for cryogenic fluids, the temperature variation can be up to
1–2 K for the hydrogen.11 The considerable temperature drop during
A. Thermodynamic properties of cryogenic fluids
the cavitation process, on the other hand, changes the local physical
Cavitation, from the standpoint of classical thermodynamics, rep- properties, which in turn affects the cavitation dynamics.12 The tem-
22 November 2024 04:35:04

resents the crossing of the saturation curve from the liquid to the vapor perature drop in the cavity together with the thermo-sensitive physical
phase as the pressure is below the vapor pressure. Hence, in most cases, properties constitutes the so-called thermal effect/thermodynamic
cavitation is a phenomena that happens at a constant temperature, with effect of cavitation in cryogenic fluids, which is the most important
only a relatively small amount of heat to produce a significant volume feature of cryogenic cavitation.
of vapor, and the flow is considered to be isothermal.11 Most of the early
studies treated cryogenic cavitation as a special case of room- B. Thermal effects evaluation
temperature fluid (e.g., water) cavitation. The isothermal hypothesis,
Understanding and predicting thermal effects in cryogenic fluids
however, is no longer reasonable in cryogenic cavitating flows for their
are crucial since the operating temperature of cryogenic fluids is close
temperature change and thermo-sensitive properties.11 Intuitively, the
to the critical point temperature.13 Quantification of the thermal
significant differences in the physical and thermal properties of cryo- effects was generally estimated in terms of a temperature drop associ-
genic fluids and water are expected to contribute to the special nature of ated with the flow conditions and liquid properties. The scale of tem-
cryogenic cavitation. The properties of common cryogenic liquids and perature difference (DT) due to the thermal effect can be estimated
water are summarized in Table I. In general, the specific differences are using a simple heat balance between the two phases as follows:
mainly reflected in the following three aspects:
qv Vv Lev ¼ ql Vl Cpl DT; (1)
(1) The saturation vapor pressure of the cryogenic fluid is more
sensitive to temperature, indicating that the saturation pressure where Cpl represents the liquid specific heat, and Lev is the latent heat
decreases quickly with the temperature depression. For of vaporization. The subscripts l and v represent the liquid and vapor
instance, the sensitivity of the saturated vapor pressure (dpv/dT) phases, respectively.
with regard to temperature for LH2 (20 K), LN2 (77 K), and
1. B-factor
TABLE I. Characteristic properties for water and common cryogenic liquids.
Based on the energy conservation relationship, the classical B-
factor theory was first formulated by Stahl and Stepanoff,14 which is
Liquid T (K) Cpl (kJ/kg K) ql/qv Lev (kJ/kg) k (W/m K) c (N/m)
defined as the volume ratio of vapor to liquid phase (Vv/Vl) involved
H2O 300 4.18 38 941.38 2437.3 0.61 0.072 in the vaporization process
LN2 78 2.04 162.44 198.34 0.14 0.0087 Vv DT qv Lev
LOX 91 1.70 235.63 212.26 0.15 0.013 B¼ ¼ ; DT ¼ : (2)
Vl DT ql Cpl
LCH4 110 3.47 265.79 513.84 0.19 0.013
LH2 21 10.14 44.66 445.52 0.10 0.0018 Here, DT is the reference temperature drop that depends only
on the liquid physical properties, and the reference temperature drops
Phys. Fluids 34, 101303 (2022); doi: 10.1063/5.0102876 34, 101303-3

DT of water, LN2, and LH2 are 0.01, 0.96, and 0.82 K, respectively. 1:0 0:8 " #0:3
v0 aref D 0:2 Dx=D
The difference in the reference temperature drop as well as the B value B ¼ Bref : (6)
v0;ref a Dref ðDx=DÞref
provides an assessment of the pronounced thermal effects in the cryo-
genic fluids, and the analysis of the B-factor has been widely used by
many investigators for the characterization of thermal effects in
b. MTWO method. The method proposed by Gelder et al.17 was
thermo-sensitive fluid cavitation.15,16 Subsequently, several other
based on the conductive heat transfer model and demonstrated good
formulations of the B-factor were developed based on the classical
predictive performance for B-factor. However, the model is inadequate
B-factor theory, as introduced in the following.
due to the convective heat transfer that ought to be involved in the
cavitation process not being included. An improved B-factor correla-
a. Gelder–Moore–Ruggeri method (Ref. 16). Typically, it is sup- tive approach with a firmer theoretical foundation was provided by
posed that the vapor volume equals the volume of the observable Hord,19 which took the simple thermal conduction model used by
cavity, but the liquid volume is poorly defined. To make the Gelder et al.17 as a particular case of the more general “convection”
B-factor theory more applicable to the development of a data corre- model they proposed. It was assumed that the “convective” heat trans-
lation approach, Gelder et al.17 and Moore and Ruggeri18 extended fer to shape-free bodies could be represented as follows:
the B-factor method by providing semi-empirical correlations that
relate several factors to a calibrated reference condition. The vol- Nux ¼ Co Rem1 Pr n1 ; (7)
ume of vapor (Vv) and liquid (Vl) are assumed to be, respectively,
proportional to the product of cavity length (Dx) and the thickness where Nu and Pr stands for the Nusselt and Prandtl numbers,
of cavity (dv) and the product of cavity length and liquid element respectively.
(dl), wherein dl was confirmed earlier to be proportional to the The characteristic thermal boundary layer thickness can be
vaporization time and square root of the product of thermal diffu- expressed in the following form:
sivity (a). With vaporization time proportional to Dx/v0 (v0 the dl ¼ C1 Rem2 Pr n2 : (8)
local flow velocity), dl is proportional to (aDx/v0)0.5, and the volume
20
ratio of vapor to liquid is, therefore, proportional to dv(v0/aDx).0.5 It was indicated by the data from Rouse and McNown and
Accordingly, the B-factor can be predicted relative to a test-derived Hord19 that the cavity interface can be mathematically described in
reference value as the form of dv ¼ C2xp, and B-factor can therefore be expressed as
22 November 2024 04:35:04

0:5 " # follows:
v0 0:5 aref 0:5 Dxref dv ðx ðx
B ¼ Bref : (3) #v
v0;ref a Dx dv;ref B¼ ¼ dv dx dl dx
#l
0 ð x 0
ðx
dv is presently assumed invariant, and the last term in Eq. (3)
becomes 1 and, therefore, be reduced. After a series of experimental ¼ C2 xp dx C1 Prn2 DxRem2
x dDx
0 0
studies for different temperatures and liquids in the venturi, the expo-
¼ C3 Dxp1 Rem2
x Pr
n2
nents in Eq. (3) were reevaluated, and the semi-empirical expression m2n2
was simplified as ql
¼ C3 an2 V0m2 Dxp1þm2 : (9)
ll
0:16
v0 0:85 aref 0:5 Dxref
B ¼ Bref : (4) C3, which is the similarity constant for the cavitating body, may now be
v0;ref a Dx
eliminated and empirical judgment made feasible by relating the pre-
According to the B-factors of the reference conditions, the unde- dicted B-factor to a reference value evaluated from experimental data,
termined B-factor of other liquids with a given temperature, velocity, E1 !E3
and cavity length can be predicted via Eq. (4). However, Eq. (4) is no aref v0 E2 Dx0 tref E4
B ¼ Bref : (10)
longer applicable when the size of the test section is different from the a v0;ref Dxref t0
referred one, and the geometric scale effect of the B-factor was, there-
fore, further experimentally studied by Moore and Ruggeri working It is apparent that Eq. (10) involves only the most common cor-
with venturi tubes of different sizes. The general equation with venturi relating parameters, and other correlating parameters including cavity
scale included was rewritten as shape formulations or Nusselt relationships could be additionally
introduced. A more physical representation of the B-factor expression
" #p
v0 n aref m D 1n Dx=D was derived by Hord19 after a series of dimensional analyses, wherein
B ¼ Bref ; (5)
v0;ref a Dref ðDx=DÞref the physical variables of cavity size, two-phase sonic velocity, and sur-
face tension were involved (Fig. 2),
where Dx/D denotes the scaled cavity length (D the characteristic E1 !E3
length) and is assumed to be constant. With more than 100 measured aref MTWO E2 Dx0
B ¼ Bref
data of Freon-114 and hydrogen from the 0.7- and 1.0-scale venturis a MTWOref Dxref
used, the values for the exponents of Eq. (5) were determined as E4 E5 E6
vref cref Dref
n ¼ 0.8, m ¼ 1.0 and p ¼ 0.3, and the complete form of Eq. (5) with ; (11)
appropriate exponents were given by v0 c D
Phys. Fluids 34, 101303 (2022); doi: 10.1063/5.0102876 34, 101303-4

where the MTWO parameter denotes the velocity ratio of the liquid where CA is the dimensionless area coefficient, Nu is the Nusselt num-
inlet velocity to the characteristic mass-limiting liquid velocity and ber, and Pe is the Peclet number.
was, in particular, designated to emphasize that this ratio is propor- Accordingly, the expression of the B-factor for the entrainment
tional to the two-phase Mach number, method can be obtained,
8 2 9 12
> q cl > DT CQ Pe qv cl 1
>
>1 þ B l >
> B¼ ¼ : (18)
< q c = DT CA Nu ql cpl DT
U1 v v
MTWO ¼ ; (12)
cl >
>
> 1þB v
q >
>
> Assuming that, for a given set of flow parameters, the volume flow rate
: ql ;
of vapor necessary to support a vaporous cavity is equivalent to the
volume flow rate of gas required to sustain a vented cavity. The flow
where cl and cv are the sound velocities of the liquid and vapor phases,
coefficient, CQ, was determined by an extensive experimental investi-
respectively.
gation in the form
c
c. Entrainment theory. The entrainment theory, first proposed Dx
by Holl et al.16 for temperature depression data correlation, has the CQ ¼ C1 Rea Fr b : (19)
D
benefit of separating the vaporization process into basic physical quan-
tities and can be applied as an alternative to the B-factor approach. The area coefficient, CA, was identified from pictures of both vaporous
Assuming that the cavitation zone is stable, the liquid is continuously and vented cavities and can be expressed as
evaporated at the interface of the cavitation zone, and its evaporation d
rate is equal to the rate of the vapor volume entrained from the cavita- Dx
CA ¼ C2 : (20)
tion zone by the wake. Vapors from liquid gasification extract heat D
from adjacent liquids and form thermal boundaries. Therefore, the fol- The empirical values in Eqs. (19) and (20) were determined through
lowing thermal equilibrium exists: the experiments and can be found in the literature in detail. The
_ v Lev ¼ ht AW ðT1 Tc Þ;
m (13) Nusselt number was then derived from Eq. (17) by using the measured
values of maximum temperature depression (DTmax) and the empiri-
22 November 2024 04:35:04

where Tc and T1 refer to the cavity temperature and the temperature cal relations for CQ and CA, with the formulation as follows:
of the liquid outside the thermal boundary, respectively. ht is the heat h
_ v denotes
transfer coefficient, AW is the interface area of the cavity. m Dx
the mass flow rate of the vapor in the cavity as follows: Nu ¼ C3 Re Pr Fr
l f
; (21)
D
_ v ¼ qv D2 U1 CQ ;
m (14) where C3, l, f and h are constants determined empirically.
Similar to the MTWO method, the entrainment theory for the B-
where D is the characteristic length, and CQ is the flow coefficient
factor is also characterized as a semi-empirical correlation formulated
given by
in terms of power-law formulas including fluid properties and flow
m_v parameters, with convective taken into consideration. The effective
CQ ¼ : (15)
qv D2 U1 factor having the form of CCQA Nu
Pe
are found in Eq. (18), which indicates
Substituting Eq. (14) into Eq. (13) yields that the B-factor based on the entrainment theory depends upon the
heat transfer parameters (Pe and Nu), as well as the entrainment and
qv D2 V1 CQ Lev area characteristics of the cavity.
DT ¼ ðT1 Tc Þ ¼ : (16)
ht AW
Equation (15) can be formulated in light of dimensionless coefficients, d. Fruman method. Fruman et al.21 considered the following two
namely, disparate flow scenarios of the two-phase cavity and full vapor cavity
CQ Pe qv Lev for estimating the B-factor, as shown in Fig. 3. A two-phase mixture of
DT ¼ ; (17) void fraction (av) and typical thickness (d) can be thought of as the
CA Nu ql cpl
cavitating region. Assuming that the two-phase mixture is entrained
with the general flow velocity, the B-factor can then be estimated as
Vv av U1 dc ; Vl ð1 av ÞU1 dc ; (22)
av
B : (23)
1 av
Equation (23) reveals that the B-factor is of order unity except for the
full vapor region, indicating the characteristic temperature difference
(DT ) can usually be considered as a reasonable estimate of the actual
temperature difference (DT).
For a full vapor cavity, only the surrounding liquid can supply
FIG. 2. Elemental model for B-factor theories. the heat for vaporization since there is no liquid inside the cavity,
Phys. Fluids 34, 101303 (2022); doi: 10.1063/5.0102876 34, 101303-5

(Cf)r is expressed as follows at very high Reynolds numbers:

17
1 x
ðCf Þr ¼ 0:006 95 ; (30)
2
where represents the equivalent roughness, which is calibrated from
Hord’s data based on the assumption that CQ ¼ 5.2 1013 and formu-
lated as ¼ 2:2Re0:5
x .
Different methods for the estimation of temperature drop based
on the B-factor theory are listed in Table II. To summarize, the tem-
perature drop during the process of cryogenic cavitation can be esti-
mated by knowing the values of the B-factor and DT. The GMR
(Gelder-Moore-Ruggeri) method, MTWO method, and entrainment
theory method are characterized as the semi-empirical correlations
that relate the specific parameters to a calibrated reference condition,
among which the MTWO method demands the most empirical
FIG. 3. Two typical cavitation cases for B-factor analysis. (a) Two-phase cavity and parameters to consider the convective heat transfer. Consequently, for
(b) full vapor cavity. the same fluid with similar geometrics as the above methods, respec-
tively, used, the corresponding method provides comparatively accu-
which has been analyzed by several investigators. According to the rate predictions. The classical B-factor method and the Fruman
thermal boundary layer effect, the B-factor can be estimated as method can otherwise be applied to obtain the temperature drop.
Nevertheless, the B-factor, on the other hand, fails to deal with
dc dc the transient (time-dependent) thermal effects because it relies on a
B¼ ¼ pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi ; (24)
dT aLc =U1 steady-state thermal balance equation.24 Additionally, the B-factor
method does not consider the sensitivity of vapor pressure to the tem-
where dc and dT are, respectively, the cavity thickness and the thermal perature depression, which contributes most to altering the cavity
22 November 2024 04:35:04

boundary layer thickness, Lc is the cavity length, and a is the thermal morphology. Therefore, despite the fact that the B-factor may provide
diffusivity. a reasonable estimate of the temperature depression, it is insufficient
Later, Fruman et al.22 developed another model to determine the for assessing the influence of the temperature depression on the cavity
temperature drop between the reference fluid and the cavity. The wall structure and the overall flow.
temperature was associated with the heat flux Q transferred by the
fluid with Pr 1, which is described by the following equation on a
2. Thermodynamic parameter
smooth plate:
Brennen25 proposed a more proper parameter for assessing the
Q ub
ðTp Tref Þs ¼ 1 ð1 PrÞ ; thermodynamic effect by incorporating it into the Rayleigh–Plesset
1 Uref (25)
2qcp Uref ðCf Þs equation for bubble dynamics,
" 2 #
where ub is the velocity at the edge of the viscous boundary layer, Cf is d2 R 3 dR
ql R 2 þ ¼ pv ðTc Þ p1 : (31)
the local friction coefficient, and the ratio ub/Uref is given by dt 2 dt
ub 2:1 Considering the pressure drop in the cavity the due to the local tem-
¼ : (26)
Uref Re0:1
x perature depression (Tc–T1), Eq. (31) can be rewritten as
" 2 #
In terms of the volumetric flow coefficient CQ, the heat flux can be rep-
d2 R 3 dR dpv
resented by ql R 2 þ þ DT ¼ pv ðT1 Þ p1 : (32)
dt 2 dt dT
Q ¼ qv Lev Vg ¼ qv Lev CQ Uref : (27)
The generation of a cavitation bubble is accompanied by heat
The temperature difference for a rough plate (Tp–Tref)r is related
and mass transfer between liquid and vapor. During its growth, the
to the one for a smooth plate according to Eckert and Drake,23
surrounding liquid provides latent heat for vaporization, and a thermal
ðCf Þs boundary layer of reduced temperature is formed around the cavita-
ðTp Tref Þr ¼ g ðTp Tref Þs ; (28) tion bubble.26 Plesset and Prosperetti27 supposed that a thermal
ðCf Þr
boundary layer with
ffi a small thickness exists surrounding the bubble,
pffiffiffiffiffi
where g (g 1) is a multiplying factor and assumed to be g ¼ 1 by with a size of al t (al the liquid thermal diffusivity), and the tempera-
Fruman22 in the model, which leads to tures on the inside and outside of the boundary layer are Tc and T1,
respectively. The schematic of the thermal boundary layer on the cavi-
qv Lev CQ 2:1
ðTp Tref Þr ¼ 1 0:1 ð1 PrÞ : tation bubble interface is illustrated in Fig. 4. According to the
1 Rex (29) Fourier’s law of heat conduction, the heat balance across the bubble
2qcp ðCf Þr interface can be expressed as follows:
Phys. Fluids 34, 101303 (2022); doi: 10.1063/5.0102876 34, 101303-6

TABLE II. Different methods for the estimation of temperature drop based on the B-factor theory.
Reference source of experimental data
Formulations of the B-factor Equation number Fluids Models Empirical parameters
GMR method Eq. (6) LH2 Venturi Velocity

Thermal diffusivity
Cavity length
Cavity thickness
MTWO method Eq. (11) LN2 and LH2 Venturi and hydrofoil Velocity
Thermal diffusivity
Cavity length
Kinematic viscosity
Surface tension
Dimension of geometry
Entrainment theory Eq. (18) Freon 113 Zero-caliber and quarter-caliber ogive Cavity length
Dimension of geometry
Fruman method Eq. (24) LN2 and LH2 Hydrofoil /
Eq. (29) R-114 Venturi
" 2 #
T1 Tc d 4 3 d2 R 3 dR pffiffi pv ðT1 Þ p1
4pR2 kl pffiffiffiffiffiffi ¼ pR qv Lev : (33) R 2 þ þ RR_ t ¼ : (35)
al t dt 3 dt 2 dt ql
Then, the temperature difference can be expressed as It can be concluded that the thermal effect on bubble dynamic directly
22 November 2024 04:35:04

pffiffi depends on the thermodynamic parameter R, which is defined as
R_ t q Lev
DT pffiffiffiffi v : (34) qv Lev dpv
al ql CPl R¼ pffiffiffiffi : (36)
q2l Cpl al dT
Substituting Eq. (34) into Eq. (32), we get
As an alternative scaling law for thermal effect assessment, the
thermodynamic parameter R avoids the primary limitations of the B-
factor by considering the time-dependent effects and the influence of a
temperature drop on the cavitation dynamics. As a result, the thermo-
dynamic parameter R has been used by many investigators to assess the
thermal effect on the cryogenic cavitation and demonstrated as an effi-
cient criterion for determining whether or not the cavitation process is
thermally controlled. Figure 5 compares the value of the thermody-
namic parameter R for water and several common cryogenic liquids.
FIG. 4. Schematic of the thermal boundary layer on the cavitation bubble interface. FIG. 5. Thermodynamic parameter R for water and several common cryogenic liquids.
Phys. Fluids 34, 101303 (2022); doi: 10.1063/5.0102876 34, 101303-7

In addition to the above two classical methods, namely B-factor within a tunnel with a glass window to record images of the cavitation
theory and thermodynamic parameter R, other forms of thermody- region. The average cavity length under a range of temperatures and
namic parameters were developed by the later researchers. For inlet velocities was well recorded, and both the temperature and pres-
instance, Francpet al.28 incorporated the non-dimensionalized R into
ffiffiffiffiffiffiffiffiffiffiffiffiffi sure were measured at five probe locations along the geometries. All
the form of R D=U1 3 (R), to include the important role of velocity
the results together with the instrumentation error were documented.
scale U1. Using a model of bubble growth in a time-varying pressure Consequently, Hord’s data are regarded as benchmark results for vali-
field, Ehrlich and Murdock29 derived a similar non-dimensional dating computational methodologies involving the thermal effects in
parameter whose expression is rtip X3=2 =RU1 3
. To assess the thermal cavitation. Figure 6 depicts the schematic of the cryogenic cavitation
effect with a thermodynamic parameter having more physical conno- experimental system employed by Hord in NASA.
tation, Chen et al.30 proposed a “C-factor” (C-factor ¼ DT dp v 1
dT 0:5q U 2
Since then, various experimental apparatus and circulating sys-
l 1
qv LeV dpv tems were developed and employed by the later investigators to study
¼ 0:5C 2 2
lq U dT ) and testified its practicality by comparing with R.
l 1 the cavitation behaviors of cryogenic liquids in different kinds of
It should be noted that predicting thermodynamic/thermal geometries. Despite limited reported data as compared to water, these
effects in cavitation is generally based on data in liquids other than experimental studies help obtain deeper insight into the mechanism
cryogenics because of the tremendous challenges in experimental and characteristics of cryogenic cavitation, and the limited but valuable
investigation. Hot water or specific refrigerants are the most often data were later used as important reference standards in terms of the
used surrogate liquids, which are chosen based on a single fluid prop- numerical models used for cryogenic cavitation.
erty, most typically the thermodynamic parameter R. However, the
general review of cavitation studies is purely restricted on the basis of A. Cavitation visualization and cavity dynamics
the relevance to cryogenic liquid since cryogenic cavitation remains
the primary interest of this study. As one of the most direct methods to characterize the cavitation
intensity as well as the features of cavitating flow regime, the visualiza-
II. EXPERIMENTAL STUDIES tion of cavitating flow field has always been the focus of experimental
Due to the extreme complexity of test facilities and the difficulties research. However, in view of the thermal insulation requirements,
of data measurements in cryogenic flows, there is a dearth of studies defrosting visualization windows, and low-temperature embrittlement
when compared with typical cavitation studies of water. Back in the of metals and plastics, it would be quite a challenge to build a test facil-
22 November 2024 04:35:04

1960s and 1970s, Hord19,31,32 conducted a series of experiments with ity that enables visualization for cryogenic fluids. Furthermore, experi-
LN2 and LH2 under varying inlet conditions and reported comprehen- mental characterization of liquid–vapor interfaces during cavitation is
sive measuring data of cryogenic cavitation in the nozzle, hydrofoil, always challenging because of the rapid evolvement of the vapor
and ogive geometries. These geometrical devices were positioned structures.
FIG. 6. Schematic diagram of the experi-

mental system for cryogenic cavitation
from NASA.19 Public use permitted from
U.S. Government.
Phys. Fluids 34, 101303 (2022); doi: 10.1063/5.0102876 34, 101303-8

Starting with the dynamic behavior recording of a single bubble, hand, can be simulated by triangle-like converging–diverging geome-
various phenomena caused by the cavitation occurring in cryogenic try. The attached leading edge cavitation of the blade can be simplified
liquid were elucidated gradually with the advancement of image as the cavitation over a hydrofoil with a pressure gradient.
recording and processing technology. To the best of the authors’
knowledge, Sato et al.33 first dealt with an experimental study on the 1. Cryogenic cavitation morphological characteristics
laser-induced bubble dynamics in LN2. The bubble is generated by
focusing and pulsing irradiation of second harmonics of the YAG laser The significant difference between Freon cavitation and water
into the cryostat, and the bubble dynamics were recorded by a high- cavitation was detected by Sarosdy and Acosta38 with the help of pho-
speed image converter camera. Not only that but the interplay of the tographs of cavities of moderate length in Freon and water. The water
rebounding bubble with the rigid wall has also been investigated. cavities were found to be clearly defined, but the Freon cavities are
Based on the results obtained, Sato’s group proposed a new laser sur- always frothy and indistinct, with many small bubbles, which has been
face processing technology to handle the cavitation bubble.34,35 substantiated to be attributed to the dominance of thermal effects in
Toshimasa et al.36 also designed a cryostat and used a digital camera Freon. Giorgi et al.11 performed experiments on LN2 and water cavi-
with a shadowgraphic method to perform a visualized experiment on tating flow in orifices and observed the temporary growth process of
the pulsed YAG laser-induced bubble behavior in LN2. The bubble the cloud cavitation by high-speed video images under different oper-
behavior along with the shock wave propagation was successfully visu- ating conditions. The experiments showed that the cavitation in LN2
alized and discussed. More recently, a series of visualization experi- has a frothy look and a considerably finer structure than in water at
ments were performed by Petkovšek and Dular37 to get a better ambient conditions, in accordance with the observations made in the
knowledge of cavitation dynamics in cryogenic liquids. Comparing the study of Sarosdy and Acosta.38 Additionally, it was observed in the
observation of cavitation in LN2 and water at different temperatures high-speed photos of water cavitation that the visual extent of cavita-
revealed that different liquids act differently while having the same tion increases with temperature since smaller pressure differences are
thermodynamic parameter R. More specifically, cavitation in LN2 is required to achieve the phase-change curve as a result of the rise in
significantly more structurally ambiguous and quickly detaches from vapor pressure. Ito et al.39 investigated the cavitation pattern of plano-
the horn tip, forming complex structures. convex hydrofoils in a LN2 high-speed cavitation tunnel. The periodic
Although the experimental investigation of the cavitation in the shedding frequency of cloud cavitation is significantly lower than that
cryogenic inducer is very important, it started very late because it is estimated by empirical formulas for vortex shedding on the surface of
22 November 2024 04:35:04

very difficult to conduct.2 Considering the complex structure of the blunt bodies. Even though the cavitation zone fully covers the hydro-
actual inducer and the coexistence and interaction of multiple cavita- foil surface, cloud cavitation still develops, but in water, cloud cavita-
tion regimes, a common approach to studying the cavitating flow tion generally happens only when the length of the cavitation zone is
regimes in the inducer is to simplify the study of each cavitation mor- shorter than the hydrofoil chord length; otherwise, it will turn into
phology to the cavitation flow regime analysis of independent simple supercavitation.
geometries that produce similar cavitation morphology, to elucidate An experimental apparatus for LN2 cavitating flow around a
the mechanism of occurrence, development, and shedding of respec- NACA66 hydrofoil with a visualized test section was built by Gao
tive cavitating flow regime. Figure 7 shows the locations of the investi- et al.40 The attached cavity length was determined by using the stan-
gated cavitating regimes in an inducer. The cavitation at the shaft end dard deviation (STD) method, which was found to be dependent upon
of the inducer can be modeled by the cavitation formed in different the cavitation number. The cavitation length, to be specific, sharply
structures of blunt-tipped bodies (right-angle, round, parabolic, etc.) increases with the decrease in cavitation number. Chen et al.41,42 con-
in the tunnel, while the internal blade passage cavitation, on the other ducted an experimental investigation of LN2 cavitating flow in the
FIG. 7. Cavitation modes over inducer.
Phys. Fluids 34, 101303 (2022); doi: 10.1063/5.0102876 34, 101303-9

convergent–divergent (C–D) nozzle with the temperature range from addition, it was also found that the cavitating flow condition is unsta-
73 to 86 K to study the temperature effects on the cavity dynamics. A ble at a low Reynolds number and is lacking in reproducibility con-
high spatial–temporal resolution high-speed camera was employed, cerning the generation of cavitation. Ohira et al.51 experimentally
and the cavity pattern, as well as the cavity area at different tempera- investigated the change of the cavitation mode of saturated LN2 and
tures under different flow conditions, could thus be compared. It was subcooled LN2 and clarified the varying instability phenomenon in a
concluded that with the increase in temperature under similar flow C–D nozzle with different throat diameters. In the experiment, the
conditions, the cavity area increases and the cavity turns frothier and LN2 was subcooled by gradually reducing the pressure. The visualized
mushier. On the basis of the clear experimental images obtained from parts are venturi tubes made of quartz glass with throat diameters of
the high-speed camera of the same experimental system, Liang et al.43 1.5 and 2.0 mm, respectively. It was experimentally observed that
analyzed the unsteady behaviors and detailed cavity structures of LN2 when the temperature of LN2 was lowered to 76 K, the flow mode
cavitating flow inside the C–D nozzle under different conditions. Four switched from continuous mode to intermittent mode. In the intermit-
typical cavitation patterns, including incipient cavity, choke cavity, tent flow mode, the time interval between the occurrence and disap-
sheet cavity and large-scale cloud cavity, were observed in LN2 cavitat- pearance of cavitation was only milliseconds.
ing flows, and the changes in cavitation patterns with temperatures The unsteady flow of LN2 via a transparent venturi tube in a vari-
were also well captured. Other studies on the visualized cavitation able pressure ratio tunnel was experimentally investigated by Zhu
regimes of cryogenic liquids through various geometrics, e.g., LHe cav- et al.52 to study the dynamic cavitation characteristics. Observations of
itating flow through a C–D flow channel,44 can be referred to in the the photographs taken in succession as the pressure ratio rises reveal
literature.45 two representative cavitation modes, namely, quasi-steady mode and
In contrast, there are few known visual experimental facilities of dynamic mode. The shedding frequency of cavitation was determined
cavitation on a rotating inducer in cryogenic liquids due to the very by applying the Fourier transform to pressure waves. It was found that
great difficulty. The initial breakthrough in this field was accomplished the maximum length of the cavity and shedding frequency both
by Ball et al.46 from NASA, who observed a helical inducer rotating in increase with the pressure ratio, while the shedding dynamics are
a big LH2 tank with a transparent cylindrical casing using a video cam- completely featured as the dynamic mode in which large-scale cavita-
era. Unfortunately, the visuals were ambiguous. Consequently, the cav- tion clouds periodically shed off when Pratio > 2.23. Later, Zhu et al.53
itation formed surrounding the inducer was found, but its mechanism performed visual experiments based on the improved facility to further
remained unclear. A few decades later, Watanabe et al.47 made a 150-
22 November 2024 04:35:04

study the LN2 cavitating flow through a venturi tube with an emphasis
mm helical inducer rotate in an LN2 cavitation tunnel and observed on the influences of thermal effects on cavitation dynamics. The pho-
the inducer with a high-speed camera through a transparent polycar- tograph of the experimental device and schematic diagram are shown
bonate cylindrical casing. The clear visualized movies revealed that the in Fig. 8. The cavity length as well as the frequency was determined
cavitation in the vicinity of the rotating inducer comprised many small from the image processing by using the standard deviation (Std)
bubbles. In addition, detailed structures that correlate with several method. The cavity length-based non-dimensional Strouhal number
types of cavitation around the inducer were identified. Almost at the (Stc) was applied to represent the characteristics of shedding frequency
same time, Ito et al.48 built the first test facility allowing the visualiza- of cloud cavitation,
tion of cavitation around a rotating inducer in both water and cryogen.
With the improvement of the experimental equipment, a series of fLc
Stc ¼ : (37)
visual images of the helical inducer in LN2 were presented, the diame- uth
ter of each backflow vortex cavitation column was measured quantita-
It turned out that the values of Stc can be used to identify the sheet cav-
tively, and the characteristics of backflow vortex cavitation and tip
itation and cloud cavitation, with Stc of 0.04–0.08 for sheet cavitation
vortex cavitation were, therefore, could be elucidated. In particular, the
and Stc of 0.30–0.40 for cloud cavitation. For cloud cavitation, the
cavitation regimes of the inducer in LN2 and water under identical
pressure ratio has a dominant effect on the shedding instability, while
flow conditions were directly compared, and the typical appearance of
collapsing backflow vortex cavitation of LN2 and water was well illus- the thermal effects determine the transition between cloud cavitation
trated in their follow-up study.49 and sheet cavitation.
2. Unsteady characteristics and flow instability B. Synchronous measurements

Cavitation is characterized by high unsteadiness, acting as one of With the continuous development of experimental techniques
the most important sources of flow instability. It is, therefore, very crit- and the deepening of cavitation research, the multiphysics synchro-
ical to perform transient cavitating flow analysis of cryogenic liquids nous measurement of cavitation flow has been more and more applied
to have a thorough understanding of the unsteady cryogenic charac- and promoted. Through the multiphysics synchronous measurement
teristics and flow field instability mechanism. Ishimoto et al.50 studied techniques, several physical quantities (such as pressure, temperature,
the fundamental characteristics of pressurized LN2 cavitating flow in a velocity, etc.) can be synchronously measured in real-time, which
horizontal rectangular nozzle, both numerically and experimentally. makes it possible to analyze the instantaneous interaction of various
Temporal evolutions of the void fraction distributions were measured flow field parameters in cavitation flow. Among the experimental syn-
by a high-speed camera. When the magnitude of the cavity exceeds a chronous measurement of cryogenic cavitation, the combination of
specific size, a cavity composed of small bubbles is observed split from high-speed flow visualization technique and pressure measurements
the main cloud cavity and advected downstream of the nozzle. In system has been extensively employed.
Phys. Fluids 34, 101303 (2022); doi: 10.1063/5.0102876 34, 101303-10

FIG. 8. Photograph of the experimental device (a) and schematic diagram (b).53 Reproduced with permission from Zhu et al. [“Influences of thermal effects on cavitation
dynamics in liquid nitrogen through venturi tube,” Phys. Fluids 32, 012105 (2020)]. Copyright 2020 AIP Publishing.
In an experimental study of cavitation unsteady characteristics of velocity of the cavitation area and the transient cavity structures. With
LN2, Chen et al.41 used a high-speed image acquisition system for cavi- this synchronous experimental system, the experimental measure-
22 November 2024 04:35:04

tation visualization and mounted three pressure sensors (measuring ments of the velocity in the cavity of cryogenic cavitation were first
range 0–0.6 MPa, accuracy 0.1%, frequency 2300 Hz) to monitor the reported, and the reentrant jet dynamic as well as the reentrant jet-
pressure upstream, downstream, and in cavity region, and high syn- dominated shedding mechanism for unsteady LN2 cloud cavitation
chronization accuracy was obtained. Based on the experimental data were illustrated. Table III summarizes the experimental studies using
obtained by the synchronous measurement system, the cavity mor- synchronous measurements in cryogenic cavitation.
phology and pressure pulsation were analyzed in-depth, and the inter- The cryogenic cavitation flow is highly unsteady, and its evolu-
action between cavity evolution and fluid pressure oscillation was tion will cause drastic changes in other physical quantities in the flow
revealed. Likewise, Chen et al.54 also synchronously investigated the field, such as pressure and velocity, which are closely related to the vor-
cavity behavior and the corresponding pressure characteristics in a tex and turbulence structure in the flow field. Therefore, the experi-
C–D nozzle with five absolute pressure transducers (measuring range mental study on the transient interaction of multiple physical
0–0.6 MPa, accuracy 60.1% FS) for the measurement of the pressure quantities in the cryogenic cavitation flow field is of great value to
along the C–D nozzle wall. Esposito55,56 studied cavitation through an reveal the cavitation mechanism, and the synchronous measurement
orifice using LN2 as the working fluid, and the dynamics of the identi- techniques of multiple physical fields will become more and more
fied cavitating regimes were described by unsteady pressure measure- important in the research of cryogenic cavitation flow. The establish-
ments and high-speed flow visualizations. ment of test facilities allowing visualization that can be utilized with
Fdida et al.57 utilized an imaging system in conjunction with laser cryogenic fluids is of great difficulty because of the thermal insulation
Doppler velocimetry (LDV) and particle tracking velocimetry (PTV) requirements, defrosting visualization windows, and low-temperature
to experimentally characterize the bubble flows in LN2. The rising embrittlement of metals and plastics.
velocity of the bubbles was determined using all three optical diagnos-
tics, and good agreement was obtained. In addition, the characteriza- III. COMPUTATIONAL MODELING APPROACHES
tion of bubble shape obtained from imaging represents valuable Despite that the experimental studies have provided rich data for
information to explain the oscillations of velocity. Dular and cavitation in cryogenic liquids and promoted the understanding of the
Petkovšek58 designed an adjustable system pressure facility to perform cryogenic cavitating flow mechanism, most experimental investiga-
erosion measurements in LN2, where cavitation was generated by an tions have been limited to obtaining pressure and/or temperature of a
ultrasonic transducer. The cavitation regime and mass loss at different limited number of monitoring points, and the accurate measurements
cavitation numbers were performed. To investigate the relationship in the cavity are more difficult to obtain. Moreover, experimental
between rotating cavitation and shaft vibration, Yoshida et al.59 study on cryogenic cavitation is inevitably time-consuming and more
applied a rotordynamic test stand and observed the growth behavior expensive, due to the extreme difficulties and stringent requirements
of cavity unevenness and synchronously analyze the rotordynamic in developing a test facility that can be employed with cryogenic fluids.
fluid forces acting on the inducer. Recently, Wei et al.60 placed an For a long time, computational modeling of cavitation has been used
LDV probe and a high-speed camera face-to-face on both sides of the to supplement experimental studies on this topic.6 Over the past deca-
test section to perform a multi-cycle synchronous recording of the des, numerical simulation has become one of the most important
Phys. Fluids 34, 101303 (2022); doi: 10.1063/5.0102876 34, 101303-11

TABLE III. Summary of experimental synchronous measurements of cryogenic cavitation.
Author Test sections Fluids Synchronous measurements
Ishimoto et al.50 Rectangular nozzle LN2 Visualization and pressure measurements

Ohira et al.51 C–D nozzle LN2 Visualization and pressure measurements
Zhu et al.52 Venturi tube LN2 Visualization, pressure and temperature measurements
Dular and Petkovšek58 Ultrasonic homogenizer LN2 Visualization and mass loss measurements
Oike et al.45 Turbopumps LN2 Visualization, pressure and vibration measurements
Gao et al.41,42 C–D nozzle LN2 Visualization and pressure measurements
Xu et al.40 Hydrofoil LN2 Visualization and pressure measurements
Esposito et al.56,61 Orifice LN2 Visualization and pressure measurements
Wei et al.60 C–D nozzle LN2 Visualization and velocity measurements
Pinho et al.62 Valve LN2 Temperature and flow rate measurements
Mena et al.63 Venturi tube LCH4 Pressure and flow rate measurements
Murakami and Harada44,64 C–D channel LHe Pressure and temperature measurements
Yoshida et al.59 Inducer LN2 Pressure and vibration measurements
Ito et al.65 Hydrofoil LN2 Visualization, pressure and temperature measurements
Fdida et al.57 Cryostat LN2 Visualization and velocity measurements
Niiyama et al.66 Orifice LN2 Pressure and temperature measurements
approaches in cryogenic cavitation research and has been applied parameter and a barotropic equation of state to include pressure
more and more widely. depression as a result of thermal effects. Ito et al.70 employed the fol-
lowing two-phase averaged state equation to describe the LN2 cavitat-
22 November 2024 04:35:04

A. Cavitation modeling ing flow over a piano-convex shaped foil:
Cavitation modeling presents considerable challenges that result 1 K ðT þ Tc Þ Rct T
from the complicated nature of this kind of two-phase flow. A compu- ¼ ð1 vÞ þ v; (38)
q p þ pc p
tational tool that includes Navier–Stokes equations, cavitation model,
and turbulence closure needs to be developed. According to how the where K, Rct, Pc, and Tc are the liquid constant coefficient, gas constant,
moving interface and multiphase mixtures are handled, various cavita- pressure constant, and temperature constant, respectively.
tion models can be classified as follows: (1) interface tracking method, On account of the missing of a cavitation transport equation, the
(2) coupled density and pressure models, and (3) transport models for coupled density and pressure model could not account for the convec-
liquid/mass fraction. tion and transport phenomena of the cavitation bubbles and is, there-
fore, better appropriate for the attached cavity rather than the
1. Cavitation models unsteady cavity.
a. Interface tracking method. The development of the interface c. Transport-equation-based model. Cavitation models based on
tracking model starts by considering a liquid–vapor interface. The transport equations are identified as a promising method to overcome
computational domains with individual phases are separated accord- the aforementioned limitations in cavitation studies. Typically, the liq-
ing to the cavity shape by time-wise grid regeneration. The pressure uid volume fraction (av) or the vapor mass fraction (fv) is determined
inside the cavity is constant, and a wake model is used to handle the by solving the transport equation as follows:
cavity closure.6 Unfortunately, this method is no longer effective as
soon as the cavity is detached. @ ðal Þ @ ðal uj Þ
þ _ þþm
¼m _ ; (39)
@t @xj
b. Coupled density and pressure model. In this method, a specific
equation will relate density and pressure in a direct way. Goncalves @ ðqm fv Þ @ qm fv uj
þ ¼m _ þþm _ ; (40)
and Patella67,68 applied two distinct liquid–vapor mixture equations of @t @xj
state (EOS) to model the cavitating flow with significant fluctuations
in the local Mach number and thermodynamic phase changes. The where the superscripts þ and –, respectively, refer to condensation
computational fluid dynamics (CFD) method is developed on the basis and evaporation. Based on the liquid–vapor mixture ratios, the fluid
of continuity, momentum, and energy conservation without any addi- property can be determined as follows:
tional cavitation models or other assumptions. The presented numeri- /m ¼ /l al þ /v ð1 al Þ: (41)
cal results demonstrate that the numerical tool is capable of properly
simulating cavitating flows on a venturi geometry. Cooper69 numeri- Due to the convective nature of the transport equation, the trans-
cally simulated LN2 pumps using a non-dimensional vaporization port-equation-based cavitation model can describe the effects of inertia
Phys. Fluids 34, 101303 (2022); doi: 10.1063/5.0102876 34, 101303-12

TABLE IV. Common cavitation models.
Cavitation model _ þ (production term)

m _ (destruction term)
m
Merkle et al.71 Cprod Maxð0; p pv Þð1 al Þ Cdest ql Minð0; p pv Þal

; Cprod ¼ 80 2 t
; Cdest ¼ 1
0:5ql U1 2 t qv 0:5ql U1
1 1
Kunz et al.72 Cprod qv a2l ð1 al Þ Cdest qv MINðp pv ; 0Þal
; Cprod ¼ 3 104 2 qt
; Cdest ¼ 1
ql t1 0:5ql U1 l 1
sffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi sffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
Schneer and Sauer73 qv ql 3 2 ðpv pÞ qv ql 3 2 ðp pv Þ
að1 aÞ að1 aÞ
q RB 3 ql q RB 3 ql
pffiffiffi sffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi pffiffiffi sffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
Singhal et al.74 Max 1:0; k fv 2 ðp pv Þ Max 1:0; k 1 fv fg 2 ðpv pÞ
Cprod ql ql Cdest ql qv
c 3 ql c 3 ql
Cprod ¼ 0:01 Cdest ¼ 0:02
Senocak and Shyy75,76 Cprod Maxð0; p pv Þð1 al Þ Cdest ql Minð0; p pv Þal
ðUv;n UI;n Þ2 ðql qv Þt1 qv ðUv;n UI;n Þ2 ðql qv Þt1
Cprod 1 Cdest 1
¼ ¼
ðql qv ÞðUv;n UI;n Þ
2
ðql qv ÞðUv;n Ul;n Þ
2 2 0:5ql U1 2
0:5ql U1
Zhang et al.77 q RTref lnðp=pcav Þ þ pcav p pcav ðTref Þexp ½p pcav =ðql RTref Þ p
Cprod l Cdest
c c
1 1
2 p pcav 2 2 pcav p 2
ql fv ; Cprod ¼ 0:01 qv 1 fv fgas ; Cdest ¼ 0:02
3 ql 3 ql
Utturkar24 ql Maxð0; p pv Þð1 a1 Þ ql Minð0; p pv Þal
q ðUm;n UI;n Þ2 ðql qv Þt1
22 November 2024 04:35:04

qþ ðUm;n UI;n Þ2 ðql qv Þt1
sffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi sffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
Zwart et al.78 3av qv 2 p pv 3anuc ð1 av Þqv 2 pv p
Cprod ; Cprod ¼ 0:01 Cdest ; Cdest ¼ 50
RB 3 ql RB 3 ql
and the detachment behaviors of bubbles. Formulation of the source empirically set as Cprod ¼ 80 and Cdest ¼ 1 for ordinary fluids like
terms shown in Eqs. (39) and (40) constitute the major effort in devel- water. For cryogenic cavitation, the constants need to be calibrated
oping the cavitation model. Different formulations of source terms and specifically determined.79
representing the production and destruction rate were proposed, as R–P equation-based method. Rather than a single bubble, the
summarized in Table IV. According to the different construction semi-empirical models are derived for bubble clusters as a whole.
methods of m _ þ and m_ , these models can be divided into three cate- In comparison, the R–P equation-based method relied on the R–P
gories, namely, the semi-empirical model, the cavitation model based equation, which describes the process of a single bubble growing or
on the Rayleigh–Plesset equation (R–P equation), and the cavitation collapsing under the action of internal and external pressure differ-
model based on interface dynamics. ences, in the form of
Semi-empirical model. Merkle et al.,71 Kunz et al.,72 and other 3c
3 2 € ¼ pv p þ pg0 R0 2c R_
researchers formulated the source terms based mostly on empirical ql R_ þ RR 4l ; (43)
judgment. The general formulation of liquid–vapor evaporation and 2 R R R
condensation rates for the transport-equation-based cavitation model where R is the bubble radius, and R0 is the initial radius of the bubble.
are, respectively, shown as follows: R_ and R€ are the first derivative and second derivative of the bubble
Cprod ð1 al Þ Maxð0; p pv Þ radius with respect to time, respectively. pv is the pressure in the bub-
_þ¼
m 2
; ble, and pg0 is the partial pressure of the non-condensable gas in the
t1 0:5ql U1 bubble. c and l are the surface tension coefficient and viscosity coeffi-
(42)
Cdest al ql Minð0; p pv Þ cient of liquid, respectively.
_¼
m 2
;
t1 qv 0:5ql U1 Ignoring the effects of the second-order term, surface ten-
sion, liquid viscosity, and non-condensable gas, the relationship
where Cprod and Cdest are the empirical constants, and t1 and U1 are between the change in the bubble radius and the pressure can be
the reference timescale and velocity scale, respectively. obtained, sffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
From a physical point of view, the influence of thermal-sensible
dR 2 jpv pj
properties is magnified for the cryogenic cavitation because the density ¼6 : (44)
ratio qql in Eq. (42) is relatively smaller. As a result, the constants are dt 3 ql
v
Phys. Fluids 34, 101303 (2022); doi: 10.1063/5.0102876 34, 101303-13

ral
obtained: sffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
The following expressionis, thus,
uv;n ¼ ~
u ~
n; ~
n¼ : (47)
d 4 3 2 jpv pj jral j
m_ ¼ 6qv pR ¼ 64pqv R2 : (45)
dt 3 3 ql
However, for cryogenic liquid, the mushy interfacial dynamic
Equation (45) establishes the relationship between the interphase model was then developed by Utturkar24 since a mushy nature of
mass transport rate and pressure during the expansion or contraction the cavity has been indicated by experimental visualizations of
of a single bubble and becomes the source term basis for multiple cavi- cryogenic cavitation. The vapor pressure is determined by the
tation models based on the mass transport equation. Based on Eq. local temperature. The cavity interface is supposed to split the
(45), Schnerr and Sauer73 proposed the first mass transport cavitation mixture and liquid areas, and the source terms derived from
model that does not require empirical constants, with only the bubble the balance condition of mass and momentum at the interface of
number density remaining to be determined. Shortly afterward the cavity are given by
Singhal et al. developed the well-known full cavitation model, with the
effects of local turbulent kinetic energy, surface tension coefficient, and ql Maxð0; p qv Þð1 al Þ
non-condensable gases included. Almost instantly, a simpler cavitation _þ¼
m ;
qþ ðUm;n UI;n Þ2 ðql qv Þt1
model was established by Zwart et al.78 that could capture the details
ql Minð0; p pv Þal
of cavitation flow very well. _¼
m ;
Interfacial dynamics models. Based on the mechanism of the cav- q ðUm;n UI;n Þ2 ðql qv Þt1
ð1a Þ (48)
itation interface dynamics, Senocak and Shyy75,76 deduced the mass ql q q
¼ l þ 1:0 l e b ;
l
transfer rate between the vapor and liquid phases and proposed the q qv qv
Interfacial dynamics models, aiming to eliminate the limitation ql q
brought by empirical constants, both for classical semi-empirical cavi- ¼ l ;
qþ qm
tation model and R–P equation-based method. Different types of
models were derived for different fluids according to the distinct flow
structures around the interface. For ordinary fluids like water, the where um,n represents the normal mixture velocity, which is deter-
Sharp interfacial dynamic model featured by a thin biphasic region mined similarly to Eq. (38). The derived model can modify the
dividing the two phases was proposed. Supposed that a hypothetical source term of an earlier model for cryogenic settings. The features
22 November 2024 04:35:04

interface is lie in the liquid–vapor mixture region, the source terms of of the interface tracking method, coupled density and pressure
this model are determined by utilizing mass and momentum balance model, and transport-equation based model are summarized as
across the cavity interface and properly eradicating the unquantifiable Table V.
terms. The physical form of the source terms is defined as
d. Nucleation-based models. The previously introduced models
þ Maxð0; p pv Þð1 al Þ are based on the assumption that both the bubble size distributions
_ ¼
m ; and number of nucleation sites always remain constant. However, an
ðuv;n uI;n Þ2 ðql qv Þt1
(46) important issue of the numerical modeling of cavitating flows is
ql Minð0; p pv Þal
_¼
m ; related to the dynamics of the vapor clusters required for nucleation,
qv ðuv;n uI;n Þ2 ðql qv Þt1 which is strongly influenced by the liquid quality and the interaction
between the vapor cavity and the turbulence.80 The nucleation-based
where uI,n represents the interface velocity, and uv,n represents the nor- models account for the coalescence, breakup, growth, and nucleation
mal velocity in the vapor phase, which can be expressed as rates of the nuclei in implementing the cavitating flow modeling.
TABLE V. Summary of the numerical methods for cavitation simulations.
Cavitation simulations Descriptions Comments
Interface tracking method Treat each phase separately by keeping the cavity Difficult to deal with the shed cavity.
pressure constant and regenerating the grid at dif-
ferent times.
Coupled density and Density and pressure are connected by the EOS Not being able to implement the convection/
pressure model or barotropic equation. transport phenomenon and other cavitation
behaviors.
Transport-equation Evaporation and condensation have separate Empirical constants that are not universal for
based model phase change rates, as well as empirical constants diverse fluids.
derived from dimensional considerations. (1) Interaction between the initial bubble radius
Apply the Rayleigh–Plesset equation to describe and simulation result. (2) Number of bubbles is
the bubble growth and collapse. assumed constant.
Interfacial dynamics can be used to substitute Calculations of the interface’s movement require
empirical constants. improvement.
Phys. Fluids 34, 101303 (2022); doi: 10.1063/5.0102876 34, 101303-14

FIG. 9. Schematic diagram of the tunnel and hydrofoil utilized in Hord’s experiments19 and the computational domain with boundary conditions.
Giorgi et al.80–82 have made great efforts in applying the nucleation- the turbulent diffusivity, G(dNUC) is the growth rate, C(dNUC, x, t) and
based method to the numerical modeling of cryogenic cavitation. B(dNUC, x, t) are, respectively, the coalescence and the breakage rates,
Based on the classical nucleation theory, the homogeneous nucleation and J(dNUC, x, t) is the nucleation rate.
rate and heterogeneous nucleation rate was, respectively, derived as The LN2 cavitating flow over a quarter caliber hydrofoil was
! numerically investigated, and results showed that the vapor cavity
2SNA 16pc3 and the nucleation region continuously expanded after applying
JHOM ¼ N exp
pMWC 3kT pV ðT Þ pl the models of nuclei coalescence and breakup, which was attrib-
! uted to more nuclei both in the wake and inside the vapor cavity,
q 2cVA 16pc3
¼ l exp ; (49) where it was also predicted that the diameter of the nuclei would
MW pMWC 3kT pV ðT Þ pl decrease.
1 !
2 1 m 2cNA 2 16pc3
JHET ¼ N 3 exp U
2 pMWC 3kT pV ðT Þ pl 2. Assessments of cavitation models in cryogenic
23 12 cavitation simulations
ql 1m 2cNA
¼ Five typical transport-equation-based models, namely, the Kunz
MW 2 pMWC
! model, Singhal model, Schnerr–Sauer model, Zwart model, and
22 November 2024 04:35:04

16pc3 dynamic cavitation model (DCM), are used to perform the computa-
exp U ; (50)
3kT pV ðT Þ pl tions of cryogenic cavitating flow around a hydrofoil, which is consis-
tent with the facility employed in Hord’s experiments.19 The geometry
where MW is the molecular weight, NA is the Avogadro number, C is of the tunnel and hydrofoil, and the corresponding computational
a coefficient which considers chemical or mechanical equilibrium, and domain with the boundary conditions are given in Fig. 9.
U is heterogeneous factor. A transport equation for the bubble number Each of the cases for LN2 and LH2 in Hord’s experiments was
was additionally introduced and solved in conjunction with the mass considered and simulated, and the respective boundary conditions are
and momentum conservation, listed in detail in Table VI. Thermal effects are considered in all the
simulations, and the temperature-dependent physical properties are
@nb @nb @u
þu þ nb ¼ JHOM þ JHET : (51) dynamically updated to introduce the thermo-sensitive behavior into
@t @x @x
the cavitation characteristics. To account for the evaporative cooling
With this model, Giorgi et al.80 investigated the cavitation phenom- effect, the energy equation is coupled with the equations for turbulent
ena of LH2 flowing in a venturi and the presented results were found cavitating flows in the cryogenic cavitation modeling framework.
agree well with experimental data. Later, they81 introduced the homoge- Calibrations of empirical constants in the model used for the modeling
neous nucleation in combination with the transport, the coalescence and of cryogenic cavitation are given in Table VII. Figures 10 and 11 com-
the breakup of the nuclei, by coupling the population balance equation/ pare the surface pressure and temperature obtained in the cases of LN2
extended quadrature-based method of moments (PBE-EQBMM) with and LH2, respectively. In addition, the maximum relative pressure
the CFD model. The PBE model can provide the prediction of the fields drop and absolute temperature drop calculated by each cavitation
of the nuclei density and the nuclei diameter, which affect directly the model are calculated as shown in Table VIII.
vapor production/destruction rates predicted by the CFD model. The
number of nuclei nb and the nuclei diameter dNUC have been introduced
as variables in the population balance equation as follows: TABLE VI. Boundary conditions for modeling the LN2 and LH2 cavitating flow over
the hydrofoil.
@nb ðdNUC ; x; t Þ
þ rx ½nb ðdNUC ; x; t ÞU
@t
Inlet Inlet Outlet
rx ½Crx nb ðdNUC ; x; t Þ Numbers in velocity temperature pressure
þ rdNUC GðdNUC Þnb ðdNUC ; x; t Þ Liquids the literature (m/s) (K) (Pa)
¼ BðdNUC ; x; t Þ þ C ðdNUC ; x; t Þ þ J ðdNUC ; x; t Þ; (52) LN2 284D 23.5 77.6 332 250
where nb(dNUC, x, t) is the number density function (NDF), U is LH2 229C 40.4 20.62 179 950
the velocity of the carrier fluid, C is the sum of the laminar and
Phys. Fluids 34, 101303 (2022); doi: 10.1063/5.0102876 34, 101303-15

TABLE VII. Calibrations of empirical constants in models for cryogenic cavitation drop are both obtained by the Schnerr–Sauer model, followed by the
modeling. Singhal model. Specifically, the low pressure regions calculated by the
Schnerr–Sauer model and Singhal model are relatively larger, with small
Cavitation Empirical constants pressure gradients in the cavity region and large pressure gradients in the
model Cv Cc selected from literatures tail of the cavity. The calculated pressure and temperature by the
Kunz 100 100 83 Schnerr–Sauer model and Singhal model are close to the experimental
results. While for 229C (LH2), the DCM model, compared with the other
Singhal 0.02 0.01 74, 77
four models, showed the maximum pressure drop and temperature drop,
Schnerr–Sauer (Nb ¼ 108) 1 1 84–86 and the calculated temperature distributions are in good agreement with
Zwart 5 0.001 87, 88 the experimental data. Nevertheless, the low pressure region (as well as the
DCM 0.02 0.01 77, 89 cavity length) is not large enough obtained from the DCM model. Both
the calculated pressure and temperature by the Singhal model, by contrast,
are in good agreement with the experimental results. To summarize, all
Different characteristics of the pressure and temperature distribu- these models calculate the mass transfer terms based on the pressure dif-
tions along the hydrofoil wall are illustrated by different models. In terms ference and have been calibrated to better predict the cavitation for cryo-
of the maximum pressure drop and temperature drop obtained by differ- genic fluids. As for the simulation of the above two working conditions
ent models for 284D (LN2), the maximum pressure drop and temperature (284D and 229C), it can be concluded that the Schnerr–Sauer model is
22 November 2024 04:35:04

FIG. 10. Pressure (a) and temperature (b) distributions along the hydrofoil surface for LN2 cavitating flow with different cavitation models.
FIG. 11. Pressure (a) and temperature (b) distributions along the hydrofoil surface for LH2 cavitating flow with different cavitation models.
Phys. Fluids 34, 101303 (2022); doi: 10.1063/5.0102876 34, 101303-16

TABLE VIII. Relative pressure drop and absolute temperature drop calculated by different cavitation models.
284D 229C
Cavitation p1 pmin p1 pmin

/(%) (%)
model p1 T1 Tmin /K p1 T1 Tmin /K
Kunz 71.08 1.22 40.01 0.86

Singhal 74.11 1.61 41.46 1.55
Schnerr–Sauer 74.45 1.68 42.21 1.10
Zwart 74.01 1.44 41.38 1.07
DCM 73.82 1.24 53.28 1.81
the most appropriate model for LN2 cavitation modeling and the qm Cl k2
lt ¼ ; Cl ¼ 0:09: (58)
Singhal model comes second, while the Singhal model shows an e
advantage in LH2 cavitation modeling, followed by the DCM and
In the above equations, the Ce1, Ce2, rk, and re are model coefficients.
Zwart model. Other simulation results of different working condi-
As for turbulent cavitating flow, however, the turbulent eddy vis-
tions from Hord’s experiments with different cavitation models are
cosity in the cavity region tends to be over-predict by the standard
presented in literatures.6,90–92
two-equation models, leading to the over-prediction of the turbulent
stresses, which significantly affects the development process of the cav-
B. Turbulence modeling ity.54 The renormalization group (RNG) k–e turbulence model and
The turbulence model also plays a crucial role in cavitation flow the realizable k–e turbulence model have been proposed and widely
simulation, in addition to the cavitation model, due to the high employed in the cryogenic cavitating flow simulation, with signifi-
unsteadiness and complex flow structures of the cavitating flow. At cantly improved capabilities in characterizing flows with acute stream-
present, the most frequently used turbulence models in cavitation sim- line curvature or vortices. Not only that, but several RNG/realizable
22 November 2024 04:35:04

ulation include the Reynolds-averaged Naiver–Stokes equation k–e based correction methods were further developed, and the descrip-
(RANS) approach (such as the two-equation turbulent model includ- tions are as follows.
ing the k–e and k–x models) and the large eddy simulation (LES).
a. Filter-based model. The filter-based model (FBM) was pro-
1. RANS-based turbulent model posed by Wu et al.93 to impose a filter on the turbulent viscosity
according to the local numerical resolution, essentially combining
RANS-based methods, in the typical form of two-equation k–e
direct numerical simulations (DNS) with RANS models. The filter
and k–x models, have been extensively utilized to describe the turbu-
function (Fl_FBM) is characterized in terms of filter size (D) as
lent cavitating flow. The standard k–e and k–x models have illustrated

good accuracy in steady-state cryogenic cavitation flow. Take, for De
example, the k and e transport equations are as follows: Fl FBM ¼ Min 1; 3 : (59)
k2
" #
@ ðqm kÞ @ qm uj k @ l @k After adding the filter, the RNG k–e model is used in the case of the
þ ¼ Pt qm e þ lþ t ; (53)
@t @xj @xj rk @xj turbulence scale is smaller than the filter size, while the turbulent vis-
" # cosity coefficient as followed is applied when the turbulence scale is
@ ðqm eÞ @ ðqm uj eÞ e e2 @ lt @e greater than the filter size:
þ ¼ Ce1 Pt Ce2 qm þ lþ ;
@t @xj k k @xj re @xj 1
lt FBM ¼ Cl qm Dj 2 ; Cl ¼ 0:09: (60)
(54)
Based on the newly developed model, Wu et al.93 observed that cavi-
where k and e represent the turbulent kinetic energy and dissipation,
tating flow over a hydrofoil exhibited significant unsteadiness. Shi and
respectively. Pt denotes the turbulence production, which is defined as
Wang91 recently used the FBM model in the numerical simulation of
@ui LN2 and LH2 cavitating flow over an ogive and 2D hydrofoil, and
Pt ¼ sij ; (55) good agreements were obtained with the experimental measurements
@xj
of Hord in terms of the pressure and temperature.
where sij is the Reynolds stress tensor, defined as
sij ¼ qm u0i u0j ; (56) b. Density-corrected model. In the FBM, the effect of the variable
! mixture density and compressibility on the turbulent viscosity in the
2qm kdij @ui @uj cavitating flow region has not been considered,94 and both can com-
qm u0i u0j ¼ lt þ : (57)
3 @xj @xi promise the accuracy of cavitating flow prediction.95 Given this, the
density-corrected model (DCM) was introduced by Coutier-Delgosha
The turbulent viscosity is given by et al.,96 allowing for the compressibility effect on cavitating flow, with
Phys. Fluids 34, 101303 (2022); doi: 10.1063/5.0102876 34, 101303-17

which the turbulence dissipative terms in the two-phase zones are measured values, indicating that the FBDCM can predict the flow
reduced. The resultant turbulent viscosity can be expressed as behavior in cryogenic cavitation. Other studies62 have also shown that
the FBDCM can capture many flow details, such as reentrant jet, cavi-
Cl qm k2 tation shedding, vortex evolution, etc.
lt DCM ¼ fl DCM ; Cl ¼ 0:09;
e
qv þ ð1 av Þn ðql qv Þ (61)
fl ¼ : 2. LES model
DCM
qv þ ð1 av Þðql qv Þ
There are always some empirical coefficients in the correction
The DCM is actually related to the vapor volume fraction (av), scheme of the RANS-based turbulent model, such as the filter size D
and it corrects the area with a large vapor volume fraction, namely, the in the FBM and the exponent n in the DCM, which inevitably deter-
cavitation area. When n ¼ 1, the DCM is reduced to the k–e turbu- mine the numerical results. Despite the higher requirements for grids
lence model. Coutier-Delgosha et al.96 first suggested n ¼ 10 for water and higher consumption of computing resources, the LES avoids rely-
at room temperature. Accordingly, the value of n was also fixed at 10 ing on such empirical coefficients and gives a good performance repre-
by Long et al.97 to model the LH2 cavitating flow. Further examination sentation of the flow field details and cavity dynamics. Therefore, the
of n in the DCM for cryogenic cavitating flows was performed by LES has gradually been widely used in cavitating flow simula-
Chen et al.54 in a convergent–divergent nozzle, which was determined tions.85,86,99,100 The numerical control equation of the LES is
according to the liquid–vapor density ratio (ql/qv) of LN2 at various !
temperatures. In general, the density-corrected model can reduce the @ @ @p @ @u i @u j @sij
ðq u i Þ þ ðq u i u j Þ ¼ þl þ ;
turbulence viscosity in the wake region and thus form unsteady flow @t m @xj m @xi @xj @xj @xi @xj
better, which has been widely used in cryogenic cavitation flow
(65)
simulation.
where sij is the subgrid-scale stress and remains an unclosed term,
c. Filter-based density correction model. Although both the FBM which is expressed as
model and the DCM model can reduce the turbulent viscosity coeffi-
sij ¼ qui uj qu i u j : (66)
cient in the flow field within a certain range, some limitations also exist
in the two models individually. The FBM model is generally used to The subgrid-scale stresses are modeled and computed as follows,
22 November 2024 04:35:04

filter and modify the vortex pulsation region of a cavitation cloud with according to the Boussinesq hypothesis:
high vapor content far away from the wall to capture the unsteady
characteristics caused by large-scale eddies, but the effect of multiphase 1
sij skk dij ¼ 2lt S ij : (67)
flow characteristics of cavitation is not considered. However, the DCM 3
model corrects the mixing density of the cavitation region, with the Three most commonly used models, the Smagorinsky–Lilly
sphere of influence mainly concentrated in the core region of the near- model,101 the wall-adapting local eddy-viscosity (WALE) model,102
wall cavitation, to account for the compressibility of the vapor–liquid and the dynamic kinetic energy subgrid-scale model,103 are presented
miscible phase. To sum up, it is different regions involving distinct to determine the turbulent viscosity in Eq. (67). In the
flow characteristics where the FBM model and the DCM model come Smagorinsky–Lilly model, the eddy-viscosity is modeled by
into play, respectively. As a result, Huang et al.98 proposed a filter-
based density correction model (FBDCM), in which different turbu- lt ¼ qL2s jS j; (68)
lence viscosity correction methods are adopted in different regions to qffiffiffiffiffiffiffiffiffiffiffiffi
form a turbulence model, to capture the interaction and dynamic where jSj 2S ij S ij , Ls is the mixing length for subgrid scales and
behavior between turbulence and cavitation, and combine the merits determined using
of FBM and DCM. The FBDCM is described as follows: Ls ¼ minðjd; Cs DÞ; (69)
Cl qm k2 where j is the von Karman constant, d is the distance to the closest
lthybrid ¼ flhybrid ; (62)
e wall, Cs is the Smagorinsky constant, and D is the local grid scale,

q q which is computed according to the volume of the computational cell
flhybrid ¼ v m flFBM þ 1 v m flDCM ; (63)
ql ql using

q C1 ðC3 qm =ql C2 Þ D ¼ V 3:
1
(70)
v m ¼ 0:5 þ tanh ð2 tanh C1 Þ: (64)
ql C4 ð1 2C2 Þ þ C2
The Smagorinsky constant in the Smagorinsky–Lilly model, Cs, is
According to Huang’s work, the DCM model was applied in the not a universal constant, which is the most serious shortcoming of this
region with large vapor content at the leading edge of the foil to repre- simple model. Nonetheless, a value of about 0.1 for Cs has been found
sent the compressibility of the attached cavity, and the FBM model is to yield the best results for a wide range of flows. To obviate the need
used to capture the large-scale cavitation vortex structure in the cavita- to specify the model constant Cs in advance, Germano et al.104 and
tion region with large vapor content at the tail of the foil. Sun et al.95 Lilly105 subsequently conceived a procedure in which the Smagorinsky
used FBDCM to improve the prediction capacity of the RNG k–e two- model constant is dynamically computed based on the information
equation turbulence model for the cavitating flow. The cavity lengths provided by the resolved scales of motion, and this method is known
of the test cases predicted by FBDCM are much close to Hord’s as “dynamic Smagorinsky–Lilly model.”
Phys. Fluids 34, 101303 (2022); doi: 10.1063/5.0102876 34, 101303-18

In the WALE model, the eddy viscosity is modeled by turbulence closure, Wang et al.86 numerically studied the transient
32 LN2 cavitating flow over a twisted hydrofoil. The horseshoe U-shape
Sdij Sdij cavity morphology evolutions were presented, and the shedding mech-
lt ¼ qL2s 5 54 ; (71) anisms of LN2 cavitation with thermal effect were discussed.
S ij S ij 2 þ Sdij Sdij In general, the LES method offers a highly accurate turbulence
modeling technique and resolves the cavitating flow over a wide vari-
where Ls and Sdij in the WALE model are defined, respectively, as ety of spatial and temporal scales, which will play an increasingly
1
important role in the future numerical studies of cryogenic cavitation.
Ls ¼ min jd; Cw V 3 ; (72)
1 1 @u C. Thermal effects modeling
i
Sdij ¼ g 2 þ g 2ji dij g 2kk ; g ij ¼ : (73)
2 ij 3 @xj Since the thermal effect is especially significant in cryogenic
liquids, the modeling for thermal effects cannot, therefore, be ignored.
Cw is the WALE constant, which has a default value of 0.5. With this
As previously stated, the evaporative cooling process regulates the
spatial operator, the WALE model is designed to return the correct
thermal effects, which is further reflected by the temperature depen-
wall asymptotic (y3) behavior for wall bounded flows. Compared with
dency of vapor pressure and physical properties. A multifarious modi-
the Smagorinsky–Lilly model, the WALE model is preferable in terms
fied cavitation model concerning the cryogenic cavitation modeling
of producing a zero turbulent viscosity rather than nonzero turbulent
will be highlighted below, in which the main efforts for the extension
viscosity for laminar shear flows, which allows the correct treatment of
of the well-understood cavitation model in isothermal cavitation in
laminar zones in the domain.
different ways were made to consider the influence of thermal effects.
The original and dynamic Smagorinsky–Lilly models, discussed
previously, are essentially algebraic models in which subgrid-scale
stresses are parameterized using the resolved velocity scales. The 1. Cavitation model modifications
underlying assumption is the local equilibrium between the transferred The initial exploration of the cryogenic cavitation model primar-
energy through the grid-filter scale and the dissipation of kinetic ily focused on the verification of the applicability of the empirical cavi-
energy at small subgrid scales. The subgrid-scale turbulence can be tation model widely utilized in isothermal cavitation simulation, and
better modeled by accounting for the transport of the subgrid-scale the empirical coefficients in these models were corrected for different
22 November 2024 04:35:04

turbulence kinetic energy. Kim and Menon103 proposed the dynamic models with different cryogenic liquids to obtain the consistency with
subgrid-scale kinetic energy model in which the subgrid-scale kinetic experimental measurements. Merkle et al.71 validated the proposed
energy is defined as cavitation model for ordinary fluids (e.g., water) and set Cprod ¼ 80.0
and Cdest ¼ 1.0 as the most desirable model parameters. However,
1 2
ksgs ¼ uk u 2k : (74) under the cryogenic condition, Utturkar24 and Hosangadi and
2
Ahuja107 stated that the previously calibrated values of the Merkle cav-
The subgrid-scale eddy viscosity, lt, is computed using ksgs as itation model (Cprod ¼ 80.0 and Cdest ¼ 1.0) are improper to agree with
1 the experiment. As a consequence, Goel et al.79 suggested Cprod ¼ 54.4
lt ¼ Ck qksgs
2
Df ; (75)
and Cdest ¼ 0.68 as optimal empirical coefficients for LN2 via numeri-
where Df is the filter-size computed from Df V 1=3 . cal experimentation. Goel et al.79 have also performed numerical
The subgrid-scale stress can then be written as experimentation for cryogenic liquids, and the constants of
Cprod ¼ 54.4 and Cdest ¼ 0.639 for LN2, and Cprod ¼ 54.4 and
2 1
Cdest ¼ 0.767 for LH2, were suggested. Tseng and Shyy108 chose the
sij qksgs dij ¼ 2Ck qksgs
2
Df S ij : (76)
3 empirical coefficients suggested by Goel et al.79 to model the cryogenic
ksgs is obtained by solving its transport equation as follows: cavitation on a hydrofoil, and great agreements with the experiment of
Hord were presented.
3 ! While for the Schnerr–Sauer cavitation model, in which no
@k sgs @u j k sgs @u i ksgs
2
@ lt @ksgs empirical parameters are introduced, the bubble number density (Nb)
q þq ¼ sij Ce q þ : (77) is still set empirically, and 1013 is generally recommended for cold
@t @xj @xj Df @xj rk @xj
water and 105–109 for cryogens.109 Zhu et al.84 successfully applied the
In the above equations, the model constants, Ck and Ce, are deter- Schnerr–Sauer cavitation with the bubble number density of 108 to
mined dynamically.103 The details of the implementation of this model obtain satisfactory results of the temperature and pressure distribu-
are given by Kim.106 tions in the cryogenic cavitation simulation, and the effective compu-
To date, the implementation of LES in the cryogenic cavitation tational strategy was then generally used for cryogenic cavitation
simulation is limited due to the expensive computing cost, but the rele- modeling.
vant published studies concerning cryogenic cavitation using the LES After a robust calibration, most of these numerical investigations
method have demonstrated an irreplaceable role in the characteriza- can predict cavitation dynamics in cryogenic fluids. However, issues
tion of cavity structures and shedding dynamics. For instance, Zhu remain that the model parameters still require adjustments for differ-
et al.100 used the LES to explore the interaction of the vortex and the ent fluids. Therefore, it is necessary to modify the cavitation model
thermal effect of LH2 cavitating flows, and the effects of reentrant from the physical essence to consider the influence of the thermal
jet and pressure waves on the cavity dynamics. Based on the LES effect.
Phys. Fluids 34, 101303 (2022); doi: 10.1063/5.0102876 34, 101303-19

a. Thermal bubble dynamics. Cavitation is typically caused by 3Cdest anuc ð1 av Þ

_¼
m
the preexisting gas nuclei carried by bulk fluids, from which mac- RB
2 ! 12 3
roscopic cavities are created and grown in regions of low pressure. 1
2 pv ðT Þ p ql cpl al
2
4 5; (83)
These nuclei, filled with air (or other gases) and vapor, are small
1 ½T1 TB ðt Þ
bubbles with a diameter of a few micrometers, and their growth 3 ql cðnÞLev qv t 2
and collapse play a crucial role in the characterization of cavity 2 ! 12 3
dynamics. Recalling the derivation of thermodynamic parameter R 1
3av qv Ccond 4 2 p pv ðT Þ ql cpl al
2
5:
proposed in Sec. I B 2, it was indicated that the thermal term in the _þ¼
m 1 ½T1 TB ðt Þ
Rayleigh–Plesset equation [Eq. (35)], in addition to the inertia

RB 3 ql cðnÞLev qv t 2
term, is very important in characterizing the bubble dynamic (84)
behaviors for cryogenic fluids. Brennen25 then defined a critical
The modified cavitation model with thermal effect was employed to
time tc to evaluate the dominant term during the growing process
model the LN2 cavitating flow over a NACA0015 hydrofoil and an
of a bubble and pointed out that if the first critical time is
ogive. The simulated time-averaged temperature drops, as well as pres-
exceeded, the bubble would present thermally controlled growth.
sure coefficients along the ogive surface, were found in good accor-
The comparisons of the approximate critical transition time from
dance with the experimental results, indicating the accuracy and
inertial to thermal growth regimes for water and common cryo-
effectiveness of the modified cavitation model.
genic liquids are presented in Table IX.109
Kim and Kim,109 meanwhile, introduced the physical corrections
From the heat balance across the bubble interface [Eq. (33)], the
to the Schnerr–Sauer model by applying an appropriate bubble growth
change rate of bubble radius can be expressed as
rate in the inertial and thermal regimes within the homogeneous mix-

dR kl @T ture framework, which successfully simulated cryogenic cavitating
¼ : (78) flows from the hydrofoil test cases with LN2 and LH2 to the more
dt qv Lev @r r¼R
practical problems of engineering such as the turbopump inducer.
Equation (78) gives a connection between (@T/@r)r¼R and R(t), In general, the majority of the numerical studies on the cryogenic
and the relation between (@T/@r)r¼R and [TB(t)T1] can be cavitation modeling started from the modification of the well-proven
obtained according to the approximation proposed by Plesset and cavitation model in cavitation by considering the thermal effects on
Zwick,
22 November 2024 04:35:04

cavitation bubble dynamics, so that the modified model can accurately
represent the cryogenic cavitation behaviors and the influence of ther-
Lev qv
R t n 2 C ðnÞ;
1
T1 TB ðt Þ ¼ 1 (79) mal effects can be studied precisely and thoroughly. Some of the modi-
ql cpl al 2
fied cavitation models are summarized in Table X.
R ¼ R t n ; (80) Alternatively, Zhang et al.77,89 calculated the variable bubble
1 ð radius based on the thermodynamic phase-change theory rather than
4n þ 1 2 1 z 3n1 dz the Rayleigh–Plesset equation. The pressure of the liquid and vapor
C ðnÞ ¼ n 1 ;
(81)
p 0 ð1 z 4nþ1 Þ 2 phases were correlated by the Young–Laplace equation, and the rela-
tion for bubble radius in evaporation and condensation and process
where R* and n are constants. were, respectively, obtained as follows:
Based on the energy balance equation introduced by Brennen,
2c
the bubble growth is described by Liu et al.94,110 as follows: Rb;e ¼ ( ) ;
pl psat ðT1 Þ (85)
1
psat ðT1 Þexp pl
dR ql cpl al2 ðq1 RT1 Þ
¼ ½T1 TB ðt Þ : (82)
dt cðnÞLev qv t 12 2c
Rb;e ¼ :
pv (86)
They then recalibrated the Zwart cavitation model with the modi- q1 RTv ln pv þ psat ðTv Þ
fied bubble growth to include the influence of the thermal effect psat ðTv Þ
given by
As a result, the dynamic cavitation model was developed, with the
evaporation and condensation rate formulated as follows:
( )
TABLE IX. Critical transition time from inertial to thermal growth regimes for water p psat ðT Þ
and common cryogenic liquids.109 Reproduced with permission from Kim and Kim, psat ðT Þexp p
“A physics-based cavitation model ranging from inertial to thermal regimes,” Int. J. ðql RT Þ
Heat Mass Transfer 181, 121991 (2021). Copyright 2021 Elsevier Ltd. R_ e ¼ Cprod
c
1
R (m/s3/2) 2 psat p 2
Fluid T (K) tc (s) qv 1 fv fgas ; (87)
3 ql
Water 305 14.22 140
p
LN2 83 8.2 104 2 10–7 ql RTln þ psat ðT Þ p 1
psat ðT Þ 2 p psat 2
LOX 90 1.6 104 5 10–6 R_ c ¼ Cdest q1 fv :
LH2 21 1.1 106 3 10–8 c 3 ql
(88)
Phys. Fluids 34, 101303 (2022); doi: 10.1063/5.0102876 34, 101303-20

TABLE X. Common modified cavitation models for cryogenic cavitation simulations.
Authors Baseline model _ þ and m

m _
0sffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi 1
Shi et al.111 Kubota model pffiffiffi
3q 2 ð p p Þ q c a Max ðT T ; 0 Þ
_ þ ¼ Cdest l @ Aav
v p 1 c
m þ l pffiffi
RB 3 ql qv Lev t
0sffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi 1
pffiffiffi
3q 2 ð p pÞ q c a Max ðT T ; 0 Þ
_ ¼ Cprod v @ Að1 av Þ
v p c 1
m þ l pffiffi
RB 3 ql qv Lev t
2sffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi sffiffiffiffiffiffiffiffiffiffiffiffi3
Sun et al.87 Zwart model
m_ þ ¼ Cdest
3a v q v 4 2 p pv ðTÞ kl 5
RB 3 ql ql Cpl t
2sffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi sffiffiffiffiffiffiffiffiffiffiffiffi3
3a ð 1 a v v 4 2 pv ðT Þ p
Þq kl 5
_ ¼ Cprod
nuc
m
RB 3 ql ql Cpl t
0sffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi 1
Chang et al.112 Zwart model
þ 3av qv @ 2 p pv 1 kl A
_ ¼ Cdest
m
R 3 q1 dt q1 Cp
0sffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi 1
3a ð 1 a Þq v@ 2 p p 1 k A
_ ¼ Cprod
nuc v v l
m
R 3 ql dt ql Cp
1=3
Kim and Kim109 Schner–Sauer model 4pð1 av Þ B ql qv av ð1 av Þ
m_þ¼3 Min R_ ; R__
i t
3av 3R3min q
1=3
4pð1 av Þgcd ql qv av ð1 av Þ
22 November 2024 04:35:04

_ ¼ 3
m Min R_ i ; R_ t
3av q
0sffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi 1
Li113 Zwart model pffiffiffiffiffi
3av qv @ 2 ðp pv ðTÞÞ q C a ð T T 1 AÞ
_ þ ¼ Cdest
pl L
m l pffiffi
RB 3 ql qv Lev t
0sffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi 1
pffiffiffiffiffi
3a ð1 a Þq v@ 2 p ð T Þ p q C a ð T 1 T Þ A
_ ¼ Cprod
nuc v v l pl L
m pffiffi
RB 3 ql qv Lev t
" #
Liu et al.94 Zwart model 3av qv Cdest 2 ðp pv ðTÞÞ 1=2 ql cpl al
1=2
þ
m_ ¼ ½T1 TB ðtÞ
RB 3 ql Lev qv t 1=2
2 !1=2 3

3Cprod anuc ð1 av Þ 4 2 pv T p ð Þ q cpl al
1=2
_¼
m l ½T1 TB ðtÞ 5
RB 3 ql Lev qv t 1=2
CFD simulations using the DCM were conducted to model the according to the B-factor. Although the model has a good consistency
LN2 and LH2 cavitating flow around NASA hydrofoil and ogive for pressure distribution of various problems, it cannot directly repro-
designs, and it was demonstrated that the DCM can perform precisely duce the temperature of the flow field.
in terms of computing temperature and pressure fields of quasi-steady Instead, a supplementary energy conservation equation besides
cryogenic cavitation. the continuity equation and momentum equation is needed to be
solved. Iga et al.115 solved the total energy conservation equation for
b. Evaporative cooling effects. As discussed above, the liquid/ the liquid phase. To evaluate the thermodynamic influence of cavita-
vapor density ratio of cryogenic fluids is relatively low, necessitating tion, a nonlinear heat source suggesting the latent heat of phase change
the vaporization of a greater quantity of liquid to sustain a cavity. was developed. This method was used to model the LN2 cavitating
Cryogenic cavitation modeling must, therefore, account for the sub- flow on the flat-plate cascade, and qualitatively similar results were
stantial evaporative cooling effects, which lead to the lowering of the obtained between the numerical simulations and experimental find-
mean fluid temperature in the cavitation area. Tsuda et al.114 proposed ings as far as the improvements in suction performance caused by the
a simple model called the “reduced critical radius model,” in which the thermodynamic effect. However, the temperature depression inside
decrease in saturated vapor pressure was assessed using a simple rela- the cavity could not be evaluated because the numerical method took
tionship between temperature drop and local vapor volume fraction no account of the heat transfer in the mixture phase. Utturkar et al.116
Phys. Fluids 34, 101303 (2022); doi: 10.1063/5.0102876 34, 101303-21

and Hosangadi et al.117 accounted for the thermal effects of cavitation Singhal et al.74 proposed a probability density function method in
by solving an additional equation for the mixture enthalpy. This which the vapor pressure can be defined as
method was widely employed by other researchers to explore the cryo-
genic cavitating flow over ogive and hydrofoil. Le et al.118 and Zhang pv ðT Þ ¼ pv þ 0:5p0turb : (91)
et al.119 introduced a simplified thermodynamic model that explicitly Here, p0turb represents the local value of turbulent pressure fluctuations,
takes latent heat into account and can deal with temperature depres- which is given by
sion. The corresponding conservation of energy can be expressed as
follows: p0turb ¼ 0:39qk; (92)
@ @ where q and k are the mixture density and local turbulence kinetic
½q C p T þ ½q Cp uj T ¼ r keff rT þ SE ; (89)
@t m @xj m energy, respectively.
The full cavitation model proposed by Singhal et al.,74 which
where SE is the energy source term, which was introduced to the cavi- accounts for the turbulent pressure, has been confirmed by Zhang
tation calculation to account for the phase change heat transfer et al.126–128 to have a reasonable prediction capability and could pro-
@ @ vide satisfactory robustness in modeling the LN2 and LH2 cavitating
SE ¼ ½q ðfv Lev Þ q uj ðfv Lev Þ : (90) flow over 2D hydrofoil and ogive. Xu et al.129,130 combined the
@t m @xj m
B-factor method and the Antoine equation to correct the local satu-
rated vapor pressure caused by latent heat transfer, and the effect of
2. Physical properties update turbulent kinetic energy on the saturated vapor pressure was also con-
sidered. The corrected expression for the saturated vapor pressure that
Temperature variations in cryogenic cavitating flow due to phase accounts for the thermal effect improves the capacity to predict the
change then have a strong impact on the temperature-dependent cavitating flows and seems to provide better precision in predicting the
vapor pressure and physical properties,120 and the spatial variations of pressure coefficient along the hydrofoil surface.
the thermodynamic properties combined with the evaporative cooling
effect are incorporated into the governing equation.121 It is, therefore,
essential to specify the temperature-dependent physical properties, D. Speed of sound
22 November 2024 04:35:04

such as vapor density (qv), specific heat (Cp), and thermal conductivity The formation of cavitation produces a mixture phase region
(kl and kv), etc., as the functions of temperature. A lookup table of where liquid and vapor coexist. The precise prediction of cavitation in
physical properties within a specific temperature range was generated the liquid–vapor multiphase mixture relies heavily on the numerical
by Utturkar,24 and the temperature-based physical property was modeling of sound propagation. The speed of sound (SoS) in this field
obtained by interpolating between the appropriate tabular entries to is significantly lower than that observed in single-phase fields.
update the physical properties from the National Institute of Consequently, an accurate and reliable description of the speed of
Standards and Technology database at the end of each computational sound is essential. By performing eigenvalue examinations on the
iteration. Chen et al.122 developed a numerical framework involving strongly conservative form of the governing equations (e.g., Wallis
the energy equation to study the effect of physical properties on LN2 speed of sound), one may obtain a closed-form of the isothermal har-
and LH2 cavitation behaviors. To clearly evaluate the influence rank of monic speed of sound (cm) in the mixture region
every physical property parameter, the surrogate-based global sensitiv-

ity assessment was conducted. The comparison results suggested that 1 av al
¼ qm þ : (93)
vapor density qv plays the most dominant role in leading to the tem- cm qv a2v ql a2l
perature drop, while the liquid viscosity (ll) slightly affects the flow
structure and the thermodynamic effects. Rapposelli and d’Agostino131 extracted the speed of sound for
Rahbarimanesh123 used the computed temperature and pressure LH2 and LOX by employing a bubble model. Ohira et al.51 proposed
fields to correct the thermophysical properties of the LNG to represent an alternate relationship between the speed of sound and vapor vol-
the baroclinic nature of the density field and temperature dependence ume fraction for cryogenic fluids formulated by Rapposelli and
of the liquid saturation properties, and the resulting fluid properties d’Agostino131 under the assumption that thermodynamic equilibrium
were well verified against the reference property database. In addition, is only reached by the total volume fractions of each phase since the
the Clausius–Clapeyron equation was applied to capture latent heat thermal contact exists only partial of the two phases during cavitation.
transfer during phase transition, and the relation between saturation The thermodynamic dependable model for the speed of sound varia-
vapor pressure and temperature is obtained. Shao et al.124 suggested a tions within a cavitating two-phase flow was given by
modified Zwart model to estimate the cavitation performance of a " " g #
cryogenic pump, and the thermodynamic effect was given strong con- ð1 aÞq p pcri
c¼ ð1 l Þ 2 þ l g
sideration. They used the fourth-order polynomial to calculate the sat- p ql cl p
urated vapor pressure of the LN2. Li et al.125 also derived a four-degree #
2
1
polynomial for the saturation vapor pressure of LN2 within the tem- aq p v
þ ð1 v Þ 2 þ : (94)
perature range from 76 to 90 K, and 90 to 122 K, respectively. p qv cv Cpv
On the other hand, several experimental studies have demon-
strated the considerable effect of turbulent fluctuations on cavitating Here, v and l represent the fraction of the vapor volume and liquid
flow. To account for the effects of turbulent pressure fluctuations, volume participating in the thermal contact, respectively.
Phys. Fluids 34, 101303 (2022); doi: 10.1063/5.0102876 34, 101303-22

In addition, Venkateswaran et al.132 suggested using perturbation pc

A1 ¼ 1 ð1 fv ÞpKl ðT þ Tc Þ ;
theory to establish an effective pre-conditioned system of equations ðp þ pc Þ2
that was accordant in both the compressible and incompressible p
regimes. It was claimed that accounting for compressibility factors A2 ¼ Rg fv þ Kl ð1 fv Þ ; (101)
p þ pc
improved cavitation dynamics. Then, Senocak and Shyy75 greatly
p T þ Tc
enhanced the implementation of the speed of sound (SoS) model into A3 ¼ Rg fv þ Kl ð1 fv Þ :
pressure-based cavitation simulations, in which two speed of sound p þ pc T
models (i.e., SoS-1 and SoS-2) were proposed and compared. As a
result, the far better capability of SoS-2 was demonstrated. On this
basis, Wu et al.93 and Utturkar et al.116 adopted SoS-2 in their respec- IV. CRITICAL RESEARCH PROGRESS OF CRYOGENIC
CAVITATION INVESTIGATIONS
tive studies for time-dependent calculations. Based on the B-factor the-
ory, Zhu et al.52 considered the thermal effects in the Wallis equation A. Thermal effects
of the speed of sound, and the consequent equation was obtained as 1. Temperature drop characteristics
" #
1 ql cpl DT 1 1 1 The temperature drop due to thermal effects, which is the defin-
¼ qm þ 2 : (95) ing feature of cryogenic cavitation, has been a major concern in cryo-
c2w qv Lev þ DTql cpl qv c2v ql c2l ql cl
genic cavitation studies. Early in the study on thermal effect
The above equation was employed to determine the speed of sound in evaluations, the temperature depressions within the cavity in LN2
the LN2 cavitation region when the maximum temperature drop DT (Refs. 31 and 32) and LH2 (Ref. 32) were experimentally measured
was obtained through the experimental measurements. The predicted and referred to as mature cavitation data for developing the correlati-
speed of sound by Eq. (95) was found to agree well with the experi- ve–predictive expressions. The B-factor and other thermal parameters
mental values. However, the modified Wallis equation has not been were introduced to quantify the thermal effects on cavitation for vari-
applied in the numerical study for cryogenic cavitation simulation. ous thermo-fluids.42 Generally, the cavitation number (r) is calculated
The speed of sound equation, which takes into account thermal from the vapor pressure at the liquid temperature at infinity (T1),
effects, has also been proposed by some other researchers. In the afore-
mentioned study by Kim and Kim,109 a barotropic equation of state p1 pv ðT1 Þ
r¼ : (102)
22 November 2024 04:35:04

2
0:5ql U1
combined with the equation of the speed of sound was used to couple
the pressure and density. The mixture speed of sound is derived from
Nevertheless, cavitation develops at the local vapor pressure depending
the following definition:
on the temperature drop under cryogenic conditions. The cavitation

@h number for cryogenic fluids (rc) can be, therefore, modified as
q
@p @T p;fv p1 pv ðTc Þ
c2 ¼ ! : (96) rc ¼ :
@q s;fv @q @h @q @h 0:5ql U12
(103)
q þ 1 q
@p T;fv @T p;fv @T p;fv @p T;fv
The two cavitation numbers r and rc can be correlated by the follow-
The mixture’s density, enthalpy, and partial derivative values are ing first-order approximation:
derived from the following mixture density and enthalpy definitions: 1 dpv
q U 2 ðrc rÞ ¼ ðT1 Tc Þ: (104)
1 ð1 fv Þ fv 2 l 1 dT
¼ þ ; (97)
q ql qv Obviously, the local temperature depression (T1–Tc) gives rise to an
increase in the effective cavitation number, and the cavitation intensity
h ¼ ð1 fv Þhl þ fv hv : (98) is reduced as a consequence.
133
Le et al. modeled the LH2 cavitating flow over a tapered hydrofoil in Yoshida et al.134 obtained the value of r and rc from the experi-
the same way, and the equation of state and the speed of sound for the mental result of cold water (without thermal effect) and LN2 (with
mixture were, respectively, given as thermal effect), respectively. Then, they related the temperature
depression vs the cavity length and concluded that the temperature
pðp þ pc Þ depression increases as the cavity develop, during which the maximum
q¼ ; (99)
Kl ð1 fv ÞpðT þ Tc Þ þ Rg fv ðp þ pc ÞT temperature depression reaches 14 K. Inevitably, the thermal effect
influences the development of the cavity length. It was elucidated by
where Rg is the constant for the vapor phase, and Tc, pc, and Kl are the Yoshida et al.135 later that the thermodynamic effect is strengthened
constants for the liquid phase, by the increase in cavity length and makes the growth of cavity length
rffiffiffiffiffiffi vffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi significantly slower. Murakami and Harada64 measured the cavitation-
dp u u A3 induced temperature depression in LHe cavitating flow, and the tem-
c¼ ¼u T
dq t A2 (100) perature depression for He I was found extremely large, as much as
A1 10% of the bulk liquid temperature, and was strongly temperature-
Cp
dependent. However, the temperature depression for He II was far
with smaller due to the superthermal conduction.
Phys. Fluids 34, 101303 (2022); doi: 10.1063/5.0102876 34, 101303-23

Niiyama et al.66 conducted experiments and numerical studies isothermal mode. Above this temperature, cavitation becomes thermo-
on temperature depression in a cavitating flow through an orifice with sensitive, which means that thermal effects are prominent and the
LN2. They observed the temperature drop at the outlet of the orifice as vapor cavity size decreases. They also found that the cavitation process
r < 0.84 and found that the temperature drop became larger as r turns into more stable when the temperature is above 85 K, with ther-
became smaller. From the experimental and numerical analysis, they mal effects completely dominating the shift in the cavitation dynamics
concluded that the thermal effects suppress and delay the development in this instance. A new thermal parameter C-factor was proposed and
of cavitation bubbles and also delay the shrink of cavity bubbles when demonstrated to be effective to evaluate the thermal effects, especially
the turbulence intensity of the cavitating flow is large. in predicting the dynamic transition from the quasi-isothermal to the
The temperature distributions around the tube wall were measured thermo-sensitive mode in LN2 cavitating flows, as shown in Fig. 12. In
by Zhu et al.52 in the experiment on LN2 cavitating flows through the addition, the study elucidated the unstable cavitation features and the
venturi tube. Four copper-constant thermocouples (T1–T4) were used to shedding process of cavitating flows in various forms of thermal
measure the temperature, and the maximum drop was found about cavitation.
1.2 K. It was noticed that the measured temperature of T3, which is put Later, Liang et al.138 took a separate view of the transition process
in the middle or just behind the cavity before it collapses, was the lowest between the quasi-isothermal and the thermo-sensitive mode, dubbing
among all the cases. Moreover, the temperature drop at T2, where the it the transition mode, in order to examine the unsteady cavitation
cavity is covered most of the time, was smaller than T3. The cause was dynamics in LH2. Three cavitation modes are separated by the transi-
attributed to the radial thermal diffusion delay from the cavity to the tion temperature as the temperature increases. Furthermore, according
wall. In addition, the relation between DT and av was discussed based to the second law of thermodynamics, the entropy production
on the B-factor theory, and it was indicated that av almost linearly approach was proposed and utilized to assess the specific thermody-
increases with the increase in DT in the LN2 cavitating flow. namic characteristics quantitatively in cavitating flows in view of the
The analysis of temperature drop characteristics based on the significant influences of heat transfer on the cavitation dynamics. The
numerical approach using the developed cavitation model with the following entropy transport equation is employed:
thermal effects in consideration has been focused on the indispensable
comparisons of predicted temperature with the measured data for @s @s @s @s
qm þu þv þw ¼ Spro;D þ Spro;C ; (105)
model validation, including LN2 (Refs. 95, 109, 127, and 136), LH2 @t @x @y @z
" !
(Refs. 95, 109, and 126), and LOX.95,109 On the other hand, the numer-
22 November 2024 04:35:04

lm @u2 @v2 @w2 @v @u 2
ical approaches are uniquely positioned for comprehensive and Spro;D ¼ 2 þ þ þ þ
T @x @y @z @x @y
detailed analysis of the temperature distribution within the cavity, and 2 2 # (106)
the influences of thermal effects on the cavitation dynamics, which @u @w @v @w
þ þ þ þ ;
will be highlighted below. @z @x @z @y
" 2 2 #
2
2. Thermal effects on cavitation dynamics km @T @T @T
Spro;C ¼ 2 þ þ ; (107)
T @x @y @z
A great many studies on the thermal effects on cryogenic cavita-
tion have been performed, with the reduction of cavity length due to where S is the specific entropy, Spro,D and Spro,C is entropy production
the temperature depression effect having been widely reported.97,137 rates caused by direct viscous dissipation and heat transfer, respec-
While thermal effects have a significant influence on the cavitation tively. From the distributions of entropy production rate, it was
dynamics of cryogenic fluids, the cavitating flow field becomes even observed that the Spro,C monotonically increases with increasing
more complicated by the presence of thermal effects.24 Therefore, the
demand has arisen to have a better understanding of the impacts of
thermal effect on cavitation dynamics.
Ohira et al.51 investigated the cavitation dynamics for both satu-
rated and subcooled LN2 and summed up the cavitation modes associ-
ated with the throat temperature. By subcooling the LN2 at the nozzle
inlet, cavitation turns intermittent and ultimately disappears. While
cavitation occurs quickly when the pressure at the throat is just below
the saturation pressure at T 78 K. In this instance, continuous cavita-
tion bubbles are formed. As the throat temperature decreases, cavita-
tion begins at considerably lower pressures than the saturated
pressure. In addition to this delay, the behavior of cavitation alters and
shorter vapor clouds develop periodically near the throat. Therefore,
the pressure fluctuations are intensified at the throat. T 74 K is the
temperature at which the flow required the greatest pressure drop to
trigger cavitation from the saturation state.
FIG. 12. C-factor for LN2 as a function of temperature.41,42 Reproduced with per-
Chen et al.41,42 offered a different categorization of cryogenic cav- mission from Chen et al., “Experimental investigation of liquid nitrogen cavitating
itation dynamics that was consistent with Ohira’s observations. At low flows in converging–diverging nozzle with special emphasis on thermal transition,”
temperatures, i.e., T < 77 K, the cavitation is characterized by a quasi- Int. J. Heat Mass Transfer 132, 618 (2019). Copyright 2019 Elsevier Ltd.
Phys. Fluids 34, 101303 (2022); doi: 10.1063/5.0102876 34, 101303-24

temperature, and the small-scale cavity is approaching the surface of the and performance of turbopumps in rocket propulsion systems.130
ogive where the magnitudes of Spro,C are significantly bigger than in Since the thermal effects have a strong impact on the cavity character-
other cavity areas in thermo-sensitive mode. Several vortices inside the istics, numerous research have been performed to study the cavitation
cavity are observed when the cavity completely detaches from the ogive. of cryogenic fluids in the inducer and analyze the influences of thermal
In subsequent studies in this series, Chen et al.54 and Liang effects. Stahl and Stepanoff14 were the first to determine the head
et al.,99 respectively, took a closer look at the dynamic instability and depression values (DHv) owing to thermal effects by making use of the
cavity structures behaviors of LN2 cavitating flow in disparate thermal B-factor method based on a quasi-static theory, in which the tempera-
cavitation modes. It was indicated that, by contrast, the pressure fluc- ture drop was assessed in terms of the vapor to liquid volume ratio.
tuations in both cavitation and non-cavitation regions become slighter They provided a diagram to analyze the net positive suction head
in the thermal mode due to the strong thermal effects. The duration of [NPSH, NPSH ¼ ðpin pv Þ=ql g] corrections for hydrocarbons based
bubble collapse in thermal cavitation is significantly longer than in on this methodology. Ruggeri and Moore139 and Kovich140 performed
inertial cavitation. The changes of cavitation patterns with temperature numerous experimental studies to obtain the cavitation performance
through the C–D nozzle in LN2 were obtained. By analyzing the influ- curves of inducers employing different working fluids and established
ences of thermal effects on different cavitation patterns, different shed- empirical correlations to estimate the impact of thermal effect on the
ding mechanisms were illustrated. cavitation performance of a pump quantitatively. Franc et al.28 mea-
To further investigate the influences of the thermal effects on sured the temperature drop of the cavity inside an inducer and derived
1=2
the unsteady cavitation, an updated experimental setup of LN2 cavi- a non-dimensional thermal parameter R (R ¼ R D=U 3 ) to
tating flow through a transparent venturi tube was built by Zhu quantify the thermodynamic effect based on visualization experiments
et al.53 to obtain a deeper insight into the influences of thermal with refrigerant R-114. The thermal effect is estimated to be stronger
effects on the length of the cavity, shedding frequencies, and transi- as R increases.
tions between the sheet and cloud cavitation. The non-dimensional Yoshida et al.47,135 studied the thermal effect of LN2 cavitation
P qffiffiffiffiffiffiffiffiffiffiffiffi
parameter C=Uth3 was introduced to quantify the impact of with temperatures of 74, 78, and 83 K inside an inducer using an indi-
rect visual approach. They discovered that the temperature drop inside
thermal effects, and the pressure ratio Pr was defined as followed to
the cavity increases with the liquid temperature, whereas the maxi-
represent the flow conditions:
mum temperature drop is limited by the triple point. The critical cavi-
pout pv ðTin Þ tation number turned to a lower level for synchronous rotating
22 November 2024 04:35:04

Pr ¼ : (108) cavitation as a result of the suppression of cavity growth due to the
pin pv ðTin Þ
thermodynamic effect, which manifested dramatically with increasing
The image processing method was applied to determine both the liquid temperature. Kikuta et al.141 conducted experiments in which
cavity length and frequency. The cavitation processes were represented the inducer rotational speed was changed in LN2 to estimate the influ-
by a series of gray images, which then be approximated using the stan- ence of rotational speed on the thermal effect. It was indicated that the
dard deviation (Std) over several hundred periods. The Std of the ele- temperature depression increases with higher rotational speed, and the
ments (i) in the gray matrix of the total n pictures was defined as degree of thermal effect becomes larger with decreasing inducer head.
vffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
! !2ffi Detailed numerical analyses have been given to quantify the ther-
u
u1 Xn X n
mal effects of cavitation and to better understand their influences on
Std ¼ t Ii;j Ii;j n ;
n j¼1 cavitation flow physics in inducers and turbopumps. The helical flat
j¼1
plate inducer of NASA was evaluated by Hosangadi et al.142 using their
CRUNCH CFDR algorithm. The findings of their simulation demon-
where Ii,j (Ii,j ¼ I Ii,j,ref) refers to the gray value corresponding to the
strated that the basic distinctions between water and LH2, and the cav-
jth picture at the ith pixel. The interface of the cavity cloud was deter-
itation performance in LH2 could be accurately predicted. Goncalves
mined by the largest time-averaged Std of those pixel points at the
et al.143 also examined another NASA helical flat plate inducer with
same axial ordinate. It was found that there is a linear inverse relation-
their FINE/TURBOTM solver, which integrates the thermal cavitation
ship between the cavity length and Pr, and the cavitation will choke the
flow in the throat below a critical value of Pr (Prc) and the cavity growth model. Results from the simulation were somewhat out of the ordi-
rate increases accordingly. Based on the values of Stc (Stc ¼ fLc =Vth ), it nary. Although the temperature dropped by around 1 K in the cavity
is possible to identify sheet cavitation and cloud cavitation. The Stc as a result of the vaporization cooling effect, the wall temperature
value for the former region falls between 0.04 and 0.08. For the latter, increased 2 K owing to the viscous heating effect. This contrasting
Stc ranges from 0.30 to 0.40. The transition from sheet cavitation to behavior may have a substantial impact on the thermal effect on cavi-
cloud cavitation is thermally sensitive and will be delayed by stronger tation in an ultrahigh-speed LH2 inducer. In their simulation, about
thermal effects. The presented results can be referred to as a benchmark 0.5–1.0K temperature depressions in the cavity were observed for a
for numerical modeling of unstable cryogenic cavitation that is cur- reference temperature of 20.55 K with an inducer tip speed of 130 m/s
rently being developed. at a rotational speed of 20 000 rpm.
Recently, great attention has been paid to the examination of the
entropy production inside the cavity since it can help understand the
3. Thermal effects on cavitation in the inducer
sources and distributions of the hydraulic loss.61 Liu et al.,110 for
and turbopump
instance, have investigated the interactions between the evolution of
The prediction and simulation of the cryogenic cavitation in the force and the production of entropy. Their simulations demonstrated
inducer/turbopump are of critical importance for the efficient design the inhibitory influence of temperatures on the thermal effect, i.e., that
Phys. Fluids 34, 101303 (2022); doi: 10.1063/5.0102876 34, 101303-25

a rise in the working temperature led to a decrease in cavitation inten- wave was earlier identified by Zhu et al.52 to be responsible for the
sity. Li et al.125 simulated the LOX flow in a turbopump to assess the dynamic shedding process of cryogenic cavitation. The most evident
influence of thermal effects on the cavitation region. The local entropy measurements of formation and the propagation of the shock wave
production indicated that the thermal effects decreased and increased were initially done on single bubble cavitation. Furthermore, the shock
the hydraulic loss on account of the velocity gradient and the tempera- wave shedding mechanism was observed by Ganesh et al.145 using the
ture gradient, respectively. x-ray densitometry and a high-speed camera, and the shock
propagation-induced shedding is shown in Fig. 13. It was found that
B. Shedding mechanisms of attached cavitation the collapse of the cavitation cloud induces a shock wave that propa-
Attached cavitation with quasi-periodic shedding behavior of gates through the flow field and travels upstream at a (super)sonic
vaporous structures is one of the several cavitation types that are most speed, albeit with some deceleration, due to the fast increase in the
closely related to engineering practice and can be formed in the low- vapor volume fraction in the cavity, which reduces the speed of sound.
pressure region of the flow section. The development of attached cavi- A large cloud cavity is then pinched off when the shock front impinges
tation is quite complex accompanied by unsteady break-off and peri- on flow at the wedge apex.
odical shedding of the large cloud cavities, causing significant Petkovšek et al.146 provided another appropriate methodology
variations in the flow field, such as low-frequency pulsation of velocity based on two high-speed cameras, with a conventional one observing
and pressure, which may give rise to flow instability. Therefore, under- the cavitating flow and determining the position of the wave and the
standing the physics of this process is, thus, of fundamental impor- other one equipped microscopic lens depicting a close-up view, to cap-
tance. The mechanism of attached cavitation shedding has been ture and determine the amplitude of the shock waves. Several exam-
studied for a long time based on various geometries, and it is acknowl- ples of shock wave dynamics, such as shock wave, emitted at cloud
edged that two mechanisms, namely, the reentrant jet shedding mech- collapse in the bulk flow (as shown in Fig. 14), beneath a vapor cloud,
anism and the shock wave shedding mechanism, govern the shedding and between the attached cavity and cavitation cloud were shown and
process. the corresponding influences on the cloud behavior were discussed. It
Knapp144 first proposed the reentrant jet shedding mechanism was also noticed that the shock wave front velocities reach more than
and stated that a near-wall jet forms at the rear of the cavity due to a 700 m/s and the maximum local pressures caused by cavitation cloud
large adverse pressure gradient and moves upstream toward cavity collapse exceed 5 MPa.
Concerning the dynamic cavitation characteristics of LN2 flow in
22 November 2024 04:35:04

shedding. Since then, numerous researchers have used experimental
and computational approaches to investigate the reentrant jet and its a venturi tube, Zhu et al.52 made a breakthrough in identifying the
critical role in cloud shedding in more depth as well as the relationship shedding mechanism of a shock wave for cryogenic cavitation. The
between cavity detachments and the reentrant jet. The reentrant jet instantaneous cavitation images were recorded by a high-speed cam-
shedding mechanism has been well established as the main cause of era, and the propagation of the condensation front of the shock wave
the detachments of cavitation clouds. However, despite the well- was well captured, as shown in Fig. 15. At first, the strong pressure
proven mechanism of the reentrant jet, the mechanism of the shock wave propagates upstream within the cavity after the vapor cloud
FIG. 13. Bubbly shock propagation-induced shedding:145 (a) 0 ms, (b) 3 ms, (c) 8 ms, (d) 19 ms, (e) 32 ms, and (f) 42 ms. Reproduced with permission from Ganesh et al.,
“Bubbly shock propagation as a mechanism for sheet-to-cloud transition of partial cavities,” J. Fluid Mech. 802, 37 (2016). Copyright 2016 Cambridge University Press.
Phys. Fluids 34, 101303 (2022); doi: 10.1063/5.0102876 34, 101303-26

segments. Finally, the cavity evolves into an unstable state and shed
off. Additionally, an extended speed of sound equation based on the
Wallis equation with thermal effect taken into account was introduced
[Eq. (88)], and the experimentally obtained propagation speeds of the
condensation front were found in good accordance with the calcula-
tions by the developed speed of the sound equation.
Later on, Jaikai et al.85 numerically studied the dynamic cavita-
tion characteristics of LH2 over a 3D NACA0015 hydrofoil, focusing
on the frequency characteristics of the isothermal/non-isothermal LH2
cavitating flow. As for isothermal LH2 cavitation simulations, it was
observed that a high-pressure wave is created immediately after the
previously detached cavity collapses downstream and then greatly
enhances the intensity of reentrant flow, prompting its movements
toward the leading edge, during which the pressure wave generated by
FIG. 14. Shock waves emitted at cloud collapse in the bulk flow.146 Reprinted with
permission from Petkovšek et al., “Visualization and measurements of shock waves wake cloud collapse plays a dominant role. While for the non-
in cavitating flow,” Exp. Therm. Fluid Sci. 119, 110215 (2020). Copyright 2020 isothermal case, a much smaller pressure wave is generated, which is
Author(s), licensed under a Creative Commons CC-BY license. attributed to both the smaller size of the shed cavity cloud and the
lower vapor content due to the thermal effect. Consequently, the pres-
sure wave cannot significantly collapse the entire attached cavity. On
collapses, and the tail part of the cavity is condensed to a lower vapor the other hand, the condensation rate m _ enhances as a result of the
content than the upstream part of the cavity when a large cavitation propagation of pressure waves within the cavity, the magnitude of
cloud collapses downstream [from Figs. 15(b) and 15(c)]. The conden- velocity divergence increases accordingly [r ~ _
u ¼ ð1=qv 1=ql Þm].
sation front then moves upstream and divides the whole cavity into The vortex dilatation term is intensified, which promotes the
22 November 2024 04:35:04
FIG. 15. Sequence visualized images of the condensation process after the vapor cloud collapse (time interval is 0.141 ms).52 (a) instant 1, (b) instant 2, (c) instant 3, (d)
instant 4, (e) instant 5, (f) instant 6, (g) instant 7, (h) instant 8, (i) instant 9, (j) instant 10, (k) instant 11, and (l) instant 12. Reproduced with permission from Zhu et al.,
“Unsteady cavitation characteristics of liquid nitrogen flows through venturi tube,” Int. J. Heat Mass Transfer 112, 544 (2017). Copyright 2017 Elsevier Ltd.
Phys. Fluids 34, 101303 (2022); doi: 10.1063/5.0102876 34, 101303-27

immediately after detachment. The attached cavity becomes short for

the transitional mode, and the strength of the reentrant jet, meanwhile,
becomes slighter. The detached cavities with mushy and frothy inter-
face slowly collapse after shedding. While for the thermal mode, the
cavitation process becomes significantly stable, the shedding mecha-
nism turns into the diffusion and dissipation process, and the attached
cavity no longer detaches by the reentrant jet.
Similarly, Liang et al.,138 who also focused on the shedding pro-
cess in different thermal cavitation modes, applied a numerical
approach based on a modified density-corrected turbulent model and
a modified transport-equation-based cavitation model with LH2 at 14,
18, and 21.33 K. At these temperatures, the modes featured by quasi-
isothermal, transition, and thermo-sensitive were, respectively, pre-
sented. Lagrangian methods were used to analyze the shedding
dynamics and the distributions of Lagrangian coherent structure
(LCS) were given, as illustrated in Fig. 17. Two Lagrangian Coherent
Structures, LCS A and LCS B, may be identified for quasi-isothermal
mode, representing the primary cavity vortex and the momentum of
the reentrant jet, respectively. LCS A is positioned roughly along the
FIG. 16. Different shedding mechanisms presented in different thermal cavitation attached cavity. The development of cloud cavitation is typically
modes of LN2 cavitating flow:42 (a) inertial mode; (b) transitional mode; and (c) ther- accompanied by a substantial reentrant jet, which results in the com-
mal mode. Reproduced with permission from Chen et al., “Unsteady characteristics
of liquid nitrogen cavitating flows in different thermal cavitation mode,” Appl. Therm.
plete detachment of the attached cavity from the ogive surface. In tran-
Eng. 156, 563 (2019). Copyright 2019 Elsevier Ltd. sition mode, the location of LCS A is below the surface of the cavity
and LCS B disappears, suggesting that the liquid can flow somewhat
formation of vortex cavitation cloud and finally separates the cavity to into the cavity as a result of the diminishing vapor volume fraction.
22 November 2024 04:35:04

several segments. The inducement of vortex for the shedding behavior The reentrant jet continues to trigger the detachment of the attached
of cryogenic cavitation was first revealed, indicating that besides the cavity. However, the strength of the reentrant jet reduces significantly.
reentrant jet and shock wave, a distinct mechanism lies in the cavi- As temperature rises to the thermo-sensitive mode, the positioning of
ty–vortex interaction leading to the shedding of cloud cavitation. A LCS A moves closer to the ogive surface, and the liquid can partially
more detailed discussion on the cavitation–vortex interaction will be flow into the cavity as a consequence of the smaller vapor volume frac-
given in Sec. IV C. tion of the cavity, which is a result of the considerable influence of
To better understand the shedding dynamics of the unsteady cav- thermal effect. The influence of the reentrant jet on the attached cavity
itating flows in LN2 with varying temperatures, Chen et al.42 built the finally disappears, and the detachment is caused by the small-scale
test setup with a C–D square nozzle and carried out experiments vortexes inside the attached cavity. Assuming the fluoroketone as the
under similar cavitation number and Reynolds number with varying surrogate for cryogenic fluids as far as the thermal effect is concerned,
temperatures from 73 to 86 K. With the increase in temperature, three Sun et al.,147 Zhang et al.,148 and Liu et al.94 also investigated the
kinds of cavitation modes, namely, the inertial mode, the transitional unsteady cavitation shedding dynamics and both highlighted the
mode, and the thermal mode were identified. In general, the unsteady strong coherent relationship of the reentrant jet and vortex structure
cavitation characteristics of various thermal cavitation modes were with the flow separation.
totally different, and they compared the shedding process and mecha- In their subsequent experimental study on the dynamic process
nisms of different modes, as shown in Fig. 16. In particular, the shed- of LN2 cavitating flows under varied free-stream circumstances, Liang
ding frequency monotonically increases as the temperature increases. et al.43 demonstrated, again, that the thermal effects considerably alter
The reentrant jet plays an important role in the inertial mode, which the shedding behaviors of the cloud cavities. Additionally, they also
triggers the detachment of the attached cavity and the cavity collapses noticed that the shedding mechanisms for sheet cavities and large-
FIG. 17. Mechanisms of the shedding dynamics in different modes.138 (a) Quasi-isothermal mode; (b) transition mode; and (c) thermo-sensitive mode. Reproduced with per-
mission from Liang et al., “Thermodynamic analysis of unsteady cavitation dynamics in liquid hydrogen,” Int. J. Heat Mass Transfer 142, 118470 (2019). Copyright 2019
Elsevier Ltd.
Phys. Fluids 34, 101303 (2022); doi: 10.1063/5.0102876 34, 101303-28

FIG. 18. Shedding mechanisms of the partially cavitation for sheet cavity (a) and large-scale cloud cavity (b).43 Reproduced with permission from Liang et al., “Experimental
investigations on transient dynamics of cryogenic cavitating flows under different free-stream conditions,” Int. J. Heat Mass Transfer 178, 121537 (2021). Copyright 2021
Elsevier Ltd.
scale cloud cavities are totally distinct. A strong reentrant jet was found the identification of the reentrant jet. Comparing the velocity varia-
to form as the attached cavity enlarges to a critical length for the large- tions at the same locations, the numerical results are in good agree-
scale cloud cavity, which ultimately causes the large-scale cloud cavity ment with the experimental data. It was observed, in both
to detach. Comparatively, the relatively small length of the attached experiments and numerical simulations, the appearance of the reen-
cavity is inadequate to form a considerable reentrant jet for the sheet trant liquid jet for all the studied cases. Both the reentrant jet and
cavity, as illustrated in Fig. 18. the condensation shocks were illustrated to govern the cloud shed-
Most recently, we combined visual experiments with the numeri- ding of LN2 cavitating flow. Take, r ¼ 0.497, for example, the typical
cal method based on the LES approach to investigate the shedding shedding cycle physically consists of two stages, i.e., the development
process of cloud cavitation of LN2 cavitating flows in a C–D square of the reentrant jet and shedding of the small-scale cavity for stage 1,
nozzle under different operating circumstances. To obtain a multi- and the collapse of detached cavities under the combined action of
cycle synchronous recording of temporary cavitation structures and reentrant jet and condensation shock propagation for stage 2, as
the velocity inside the cavitation region, a high-speed camera and laser seen in Fig. 20. The evolutions of the reentrant jet, as well as the
22 November 2024 04:35:04

Doppler velocimetry (LDV) were implemented. A photograph of the shock wave propagation associated with the cavity dynamics, were
experimental device and the layout scheme of the flow field visualiza- well examined, and it was indicated that when r ¼ 0.497, the reen-
tion experiment and velocity measurement system are shown in trant jet is the predominant shedding mechanism of cloud cavita-
Fig. 19. Based on the image postprocessing approach and the experi- tion, whereas the condensation shocks instead dominate the
mental data, the distinctive shedding characteristics at various cavita- shedding process when r ¼ 0.386. The shedding mechanisms eluci-
tion numbers were analyzed. dated by the experimental and numerical investigations give a
It was reported for the first time that the velocity in the cavita- deeper knowledge of the unsteady shedding of cloud cavitation in
tion regions for cryogenic fluids was measured, which is crucial for cryogenic fluids.
FIG. 19. Photograph of the experimental device (a) and layout scheme of the flow field visualization and velocity measurement experiment system (b).60 Reproduced with per-
mission from “Unsteady cloud cavitation mechanisms of liquid nitrogen in convergent–divergent nozzle,” Phys. Fluids 33, 092116 (2021). Copyright 2021 AIP Publishing.
Phys. Fluids 34, 101303 (2022); doi: 10.1063/5.0102876 34, 101303-29

22 November 2024 04:35:04

FIG. 20. Contours of the evolution of VN2 volume fraction (a) and streamwise velocity (b) during the cavitation shedding process (the white isoline denotes av ¼ 0.1, and the
red arrow denotes the upstream moving shock front).60 Reproduced with permission from “Unsteady cloud cavitation mechanisms of liquid nitrogen in convergent–divergent
nozzle,” Phys. Fluids 33, 092116 (2021). Copyright 2021 AIP Publishing.
C. Interaction of cavitation and vortex Here, ~x is vorticity, and ~ u is velocity. The left side of the equation
Unsteady flow behaviors and periodic shedding of attached cavi- reflects the rate of vorticity change, whereas the four terms on the right
tation not only cause drastic changes in cavitation cloud structures but side are, in order, the vortex stretching term, vortex dilatation term,
also significantly enhance the local vortex distribution, which has baroclinic torque term, and viscous diffusion term. Different physical
important effects on the evolution of vortex structure. To further interpretations are implicated by these terms as followed. The vortex
investigate the cavitation–vortex interaction, Ji et al.149–151 introduced stretching term describes the stretching and deformation of the vortex
the vorticity transport equation with a variable density in the analysis structure by the cause of the velocity gradient, the vortex dilatation
of cavitation flow, term implies the stretching of vorticity produced by the flow com-
pressibility, and the baroclinic torque term represents the variations in
x
D~ rqm rp
¼ ðx
~ rÞ~ ~ ðr ~
u x uÞ þ ~ : (109)
þ ðvm þ vt Þr2 x the vorticity owing to the intersection of density and pressure surfaces,
Dt q2m and the viscous diffusion term refers to the diffusion of vorticity
Phys. Fluids 34, 101303 (2022); doi: 10.1063/5.0102876 34, 101303-30

caused by the viscous effects. It should be noted that the viscous dissi- ql qv
_ ¼
m r ~
u: (111)
pation term is small compared to the first three terms and can be ql qv
ignored in most cases. It is inferred that, from Eq. (111), the vortex dilatation term is propor-
Based on the vorticity transport equation, Zhu et al.100 ana- tional to the mass transport velocity between phases. Consequently,
lyzed the dynamic interactions between vortices, cavitation, and the vorticity distribution for the isothermal case presented by Zhu
thermal effects in partially shedding mode (PSD) and fully shed-
et al.100 could be explained in terms of the full phase change area con-
ding mode (FSM) of LH2 cavitation. The LH2 cavitating flow over
centrated at the liquid–vapor interface. For non-isothermal cavitation,
an ogive was numerically studied in a compressible mathematic
in addition to the contribution of the vortex dilatation term in the cav-
framework using the LES approach in conjunction with the
ity, the liquid-containing mixture fluid enters the cavity and interacts
Schnerr–Sauer cavitation model. The features of cavity and vorticity
with the near-wall reentrant jet at the leading edge of the ogive, giving
under isothermal and non-isothermal conditions were compared
rise to the pressure gradient, which increases the value of baroclinic
both in PSM and FSM, and the comparisons of the vorticity fea-
torque term. As a result, the vortex is further strengthened.
tures in PSM are shown in Fig. 21. It was found that strong vor-
Long et al.97 also dealt with the unsteady LH2 cavitating flow
tex–cavitation interactions in the vicinity of the ogive surface and
over an ogive with thermal effects in consideration. They also studied
in the small shedding vapor clouds were presented for the non-
the cavitation–vortex interactions based on the vorticity transport
isothermal case, while completely different vorticity distributions
equation, and the corresponding quantities were defined as follows:
were shown for the isothermal case.
To illustrate the reasons for the significant difference between the "
Dxz rq rp
two, Zhu et al.100 analyzed the influence of the thermal effect on cavity ¼ xðr ~ uÞ z þ ; (112)
behaviors. For isothermal cavitation, such as water at room tempera- Dt q2 z
ture, it is assumed that the incoming liquid can only flow around the @uy @ux
cavity but cannot enter the interior, and the liquid fully evaporates at xz ¼ ; (113)
@x @y
the liquid–vapor interface. This widely accepted assumption properly
@ux @uy
describes the features of water cavitation. While for non-isothermal ½xðr ~
u Þ z ¼ xz þ ; (114)
cavitation, it was specifically observed that the liquid can flow across @x @y

the liquid–vapor interface and flow into the cavity and cavitate (see rq rp 1 @q @p @q @p
¼ :
22 November 2024 04:35:04

Fig. 21), for the reason that the local saturated vapor pressure in the (115)
q2 z q @x @y @y @x
cavity decreases due to the thermal effect, and the difference between
the local pressure and the saturated vapor pressure becomes smaller The predicted dilatation term and baroclinic torque term distri-
accordingly so that the liquid vaporization process is inhibited and the butions were depicted, and the dilatation term was found mainly dis-
partially evaporated liquid can thus flow into the cavities. For cavita- tributed along the ogive surface, while the baroclinic torque term
tion flow, the transport equation for the vapor volume fraction can be generally concentrates along with the interfaces of the detaching cavi-
expressed as ties or near the ogive surface, with a smaller intensity than the vortex
dilatation term. They also demonstrated the predominant role of baro-
dav @av qm
¼ þ ð~
u rÞav ¼ r ~
u (110) clinic torque in vorticity production and accounted for vorticity varia-
dt @t ql qv tions along with the interface of the primary detached cavity, which
and the mass transfer rate can be obtained as follows: indicates that both the baroclinic torque term and the dilatation term
are predominant in the cavitation–vortex interactions.
Rahbarimanesh et al.121,123 and Wang et al.86 subsequently studied
the cavitation–vortex interaction for LNG cavitating flow in a nozzle
and LN2 cavitating flow over a twisted hydrofoil, respectively. Both the
different features for isothermal and non-isothermal cases and the rela-
tion between the instantaneous cavity and the vorticity field were ana-
lyzed. In their respective studies, Rahbarimanesh et al.121,123 observed
large nonzero values near the liquid–vapor interfaces for baroclinic tor-
que term, which they regarded to be caused by the density gradient due
to the temperature variations. Wang et al.86 carefully contrasted the
shedding mechanism of non-isothermal with the isothermal case, and
they demonstrated that it is the interactions of the vortex and cavity,
rather than the reentrant jet, that mostly governs the shedding process.
In brief, the cavity cloud is formed when the intensity of the vortex
inside the cavity increases to the point that it can tear off a small piece of
the cavity and then sweep away the majority of the cavity.
D. Cryogenic cavitation-induced vibration/noise

FIG. 21. Mechanisms of the unsteady cavities in LH2 with (a) and without thermal Typically, three stages are involved in the quasi-periodical sheet/
effects (b). cloud cavitation, which are the growth of the attached cavity, the
Phys. Fluids 34, 101303 (2022); doi: 10.1063/5.0102876 34, 101303-31

FIG. 22. Schematic of flow around cavity closure depicting the traveling small-scale cavities (a), shock wave emission (b), and collapsing bubble (c).
shedding of the attached cavity, and the development and collapse of depicted in Yamoto’s work. From the graph, the deviation of the bub-
the detached cavities. With the shed vapor cloud collapsing down- ble deformation to the wall was observed. Also, it was found that the
stream, bubbles experience complex nonlinear oscillation processes bubble is expanding spherically and starts to shrink at about 60 ls,
such as growth, collapse, rebound, and re-collapse. When the cavita- then the bubble subsequently deforms, collides with the wall surface,
22 November 2024 04:35:04

tion collapses, a large instantaneous pressure will be generated, causing and ultimately collapses. Similar bubble behavior in the former case
the cavitation damage to the fluid mechanical device and leading to was observed for a larger distance of c ¼ 2.01. However, due to the sig-
intense noise and vibration, among which the cavitation-induced nificant differences in physical properties between the LN2 and water,
vibration and noise have attracted the most attention and progressively distinct bubble behaviors in LN2 and water were observed after the
became a well-studied issue in terms of water cavitation. The sche- first rebounding of the bubble and following the oscillation phase. It
matics of flow around cavity closure depicting the traveling small-scale was then claimed by Yamamoto et al.152 that the cavitation bubble
cavities, shock wave emission, and collapsing bubble are illustrated in generated in LH2 or LOX served as fuels in aerospace is distinct from
Fig. 22. Despite that the vibration and noise in cryogenic fluid cavita- that in water, which demands further research work. Additionally, it
tion have great influence in many engineering fields including aero- was also realized by Yamamoto et al.152 that there is an uncommon
space engineering, it remains a related little research and poorly shape of bubble bottoms for both LN2 and water at the rebound stage,
understood. With the extensive and in-depth research on cryogenic which is known as the liquid microjet induced from the non-spherical
cavitation, the cryogenic cavitating flow-induced vibration and noise collapse of the bubbles, but a deeper inquiry was hard to come by due
have been gradually a concern by researchers in recent years. to the limitations of experimental research methods.
Recently, Mishra et al.153 numerically studied a cavitating bubble
1. Near-wall bubble dynamics and shock wave collapsing near a curved rigid surface dipped in LN2 to examine the
emission influence of varied surface configurations, with an emphasis on the
By paying attention to the fact that actual cavitation occurs in the different dynamic features such as the bubble shape evolution and jet
vicinity of solid surfaces in pumps, turbines, etc., the pilot studies on formation for collapsing LN2 cavitating bubble. A VOF interface cap-
the collapse of bubbles near or attached to a solid wall, ahead of the turing method based on the single fluid homogeneous mixture model
cavitation damage, were fundamental. The experimental inquiry on was used in a compressible framework, and the Tait equation and ideal
the near-wall bubble dynamic behavior of cryogenic liquid, pioneered gas equation were used for the modeling of compressibility of the liq-
by Watanabe et al.34 and Yamamoto et al.,152 used Nd: YAG laser uid and vapor phase, respectively. By comparing the bubble motion
beam to produce a single bubble in LN2 and distilled water near a rigid and liquid jet formation near boundaries of different configurations, it
wall in the cryostat. The laser-induced single bubble in the experiment was noticed that the bubble period near convex surfaces is shorter
was captured by the high-speed imaging system and a digital still cam- than that near-flat surface, and the maximum jet velocity at the bubble
era. A non-dimensional distance c defined by the distance from the center convex surfaces is lower than that with a flat surface. Moreover,
focusing point to the solid wall surface (c ¼ h/Rmax), along with the it has been also noticed that for cs ¼ 0.17 (cs dimensionless stand-off
non-dimensional bubble position (top, bottom) was applied to repre- distance from the surface to bubble center), the lateral parts of the bub-
sent the influence of the wall effect on cavitation bubble dynamics. ble collapse more quickly than the far end, which leads to necking and
The visualized evolutions of bubble behavior under the pressur- splitting of the bubble similar to the experimental observation of
ized conditions (þ196 kPa) in LN2 in the case of c ¼ 1.41 were Watanabe et al.34 and Yamamoto et al.152 One part repels from the
Phys. Fluids 34, 101303 (2022); doi: 10.1063/5.0102876 34, 101303-32

convex surface during the expansion that followed, while the other on the inducer is generated by the non-uniform velocity and pressure
develops a rapid reentrant jet from where the splitting occurs that distributions around the inducer due to the uneven cavity length,
directly points toward the convex surface. which will cause rotor displacement and whirl motion. The whirl
In addition to the formation of the microliquid jet, the shock motion, on the other hand, generates a nonuniform tip clearance and
wave induced by the collapse of bubbles serves as another primary accelerates unevenness due to the leakage flow in the nonuniform tip
contribution to the erosive strength of collapsing bubble. Toshimasa clearance.
et al.36 observed, by employing a digital still camera with the schlieren Shimura et al.155 and Oike et al.45 also experimentally studied the
method, the initial behavior and subsequent motion of a laser-induced LN2 rotating cavitation and its induced vibration of inducers, and they
cavitation bubble near a rigid wall and the generation of shock waves have taken it a step further to improve the design of the studied cryo-
at the moment of the bubble collapse. The shadowgraphs of shock genic inducer to suppress the rotating cavitation and attenuate the
wave propagation with vapor bubble generation for atmospheric con- effects of vibration. In the aspect of numerical simulations, studies
ditions and pressurized conditions, as well as the shadowgraphs of the have been concentrated on cavitation performance,125,156,157 flow
shock wave passing the laser-induced bubble were well illustrated. The instability,7,8,158 and cavitation suppression.130 However, the cavitating
experiment results show that the shock wave propagation patterns of flow-induced vibrations were rarely addressed due to limitations in
pressurized conditions are similar to that of atmospheric pressure con- terms of grid technique. Based on the dynamic mesh method, the
ditions, and the speeds generally agree with the speed of sound in LN2 modeling of flow-induced vibration (FIV) was well addressed in our
of 77 K, which is around 855 m/s. From the observation of shock wave recent numerical study on LN2 cavitating flow-induced vibration of a
and bubble interaction, it was confirmed that the shock waves emitted poppet-type control valve.5 The disk vibration was implemented by
from the bubble development reflect the walls and again interact with applying the dynamic mesh method based on the fluid force and stem
the laser-induced bubble, which was summed up by Toshimasa et al.36 resistance determined in the user-defined functions (UDF), and the
as “shock wave reflecting from bubble” and “shock wave passing the vibration characteristics of the valve disk were presented. Meanwhile,
bubble.” the temporal behaviors of cavitation coming with the vibration were
also captured, as shown in Fig. 23.
2. Cavitation instability and cavitating flow-induced By analyzing the evolutions of pressure distributions on the surface
vibrations of the valve disk, it was found that the pressure distribution is always
asymmetric, resulting in an uneven force at the valve’s natural position.
22 November 2024 04:35:04

Cavitation instabilities are one of the major concerns for high- At a valve opening of 30%, the valve disk was shown to vibrate periodi-
speed turbomachinery, such as turbopumps for rocket engines, cally with a frequency of about 116.6 Hz. Comparisons of non-
because these instabilities can cause structural vibrations and damage cavitation and cavitation conditions were also made, and it was indi-
to the launch vehicle elements. Therefore, it becomes extremely cated by the spectral analysis of the pressure fluctuations that the pres-
important to describe the physical mechanisms underlying this ence of cavitation promotes temporary asymmetric loads on the valve
regime. Essentially, cavitation dynamic instabilities are caused by the disk and causes immediate enormous pressure fluctuations in the flow
mass-storing capacity of compressible vapor bubbles, which delay sys- field. The developed numerical methodology for the FIV study of the
tem response to disturbance mass flow oscillations and feed excess cryogenic valve can not only help provide great insight into the mecha-
mechanical energy into the otherwise dissipative system. nisms of the complex cavitation–vibration interaction inside the valve
Rotating cavitation in inducers is a kind of cavitation instability disk, but also offers an applicable solution to model the complex
where an asymmetrical cavity pattern propagates in the same direction unsteady cavitating flow associated with vibration in the inducer.
as the rotor. In an early experimental study conducted by Yoshida
et al.,47 rotating cavitation in the inducer with LN2 of different temper-
3. Cavitating flow-induced noise
atures was investigated, and the shaft displacement sensors were fixed
at the backend of the rotor to study the vibration characteristics of the Apart from vibration, the continual bubbles generation and col-
inducer shaft. The variations of cavity length, fluid force, and shaft lapse characterizing the unsteady cavitating flow goes along with noise
vibration were recorded, and it was noticed that the fluid force acting either.159,160 Compared to a good few studies on cryogenic cavitation
on the inducer considerably goes up under supersynchronous/syn- instability and cryogenic cavitating flow-induced vibrations, very little
chronous rotating cavitation. To have a greater grasp of the link attention has been devoted to cryogenic cavitating flow-induced noise,
between the rotating cavitation and shaft vibration, Yoshida et al.59,154 neither experimentally nor numerically. Exploratory in this endeavor,
observed the growth of cavity unevenness and shaft vibration of a the unsteady LN2 cavitating flow and its induced noise around a 3D
three-bladed inducer later. From the experimental observations, it was hydrofoil, were numerically investigated in our new study.161
inferred that the shaft vibration caused by the rotating cavitation is In the study, the LES in conjunction with the Schnerr–Sauer cavi-
one sort of self-excited vibration, that is, associated hydrodynamics tation model and the permeable Ffowcs Williams–Hawkings (FW-
with rotordynamics with respect to the nonuniform tip clearance (tip Hpds) acoustic analogy approach was presented. The numerical
leakage vortex cavitation). The hydrodynamics–rotordynamics cou- method was validated, and the instantaneous vapor structures and
pling mechanism was elucidated. They provide an explanation that the pressure distributions predicted by the numerical simulations were
cavitation instability develops as flow field dynamics when the cavity found to be in satisfactory agreement with the experimental data.
length approaches a threshold length, and then the three blades’ iden- Upon comparing the general sound field characteristics of the cavitat-
tical cavity lengths become unequal, indicating the commencement of ing and non-cavitating flow, the overall sound pressure level (OSPL)
synchronized rotating cavitation. Consequently, the fluid force acting of the cavitating flow was found significantly higher than that of the
Phys. Fluids 34, 101303 (2022); doi: 10.1063/5.0102876 34, 101303-33

22 November 2024 04:35:04

FIG. 23. Contours of VN2 volume fraction evolutions during one oscillation cycle (T):5 (a) T/6, (b) T/3, (c) T/2, (d) 2T/3, (e) 5T/6, and (f) T. Reproduced with permission from
Wei et al., “Computational fluid dynamics analysis on flow-induced vibration of a cryogenic poppet valve in consideration of cavitation effect,” J. Zhejiang Univ. Sci. A 23, 83
(2022). Copyright 2022 Zhejiang University Press.
non-cavitating flow, which was mainly attributed to the contribution of applications, such as aerospace systems, chemical engineering,
cavitation noise. In particular, the time trace of the noise signal as well nuclear physics, and so on. The understanding of the cavitating flow
as the acoustic spectra was utilized to obtain a deeper understanding of characteristics in cryogenic liquid is crucial for applications like tur-
the characteristics of cavitation-induced noise. The mechanism of the bopumps for LH2 and LOX in the propulsion systems of liquid
violent fluctuations and dramatic rise of noise level during the shedding rocket engines. The complexity of the cavitation phenomenon
process of the cavity was detailed analyzed. The contours of the pressure becomes more pronounced in the case of cryogenic fluids due to the
field depicting the collapse of a detached cloud are shown in Fig. 24. substantial thermal effects and strong variations in fluid properties,
For both sheet and cloud cavitation, the collapse of the shed which significantly impact the cavitation bubble dynamics, cavity
small-scale cloud cavity is the primary mechanism that generates morphology, flow instability, frequency characteristics, etc. The pre-
intense acoustic impulses. However, the strong interactions between sent paper provides a comprehensive overview of the progress made
the reentrant jet and the main flow give rise to the violent pressure in cryogenic cavitation. The early analysis of the thermal effect has
oscillations and, therefore, create immediate extreme dipole noise, mostly been focused on deriving correlations for temperature drop
which accounts for another special mechanism for cloud cavitation to as a function of liquid properties and flow conditions. Plentiful
generate an intense acoustic impulse. experimental and simulation studies have subsequently been carried
out, which promote the understanding of the characteristics and
V. SUMMARY AND OUTLOOK mechanisms of cryogenic cavitation. Statistically, these experiments
Cryogenic cavitation has been attracting a great deal of atten- and simulation studies are comprehensively examined, as illustrated
tion with cryogenic fluids being used in a wide variety of engineering in Fig. 25, including which cryogenic working medium is used to be
Phys. Fluids 34, 101303 (2022); doi: 10.1063/5.0102876 34, 101303-34

(1) Due to the complexity of test facilities and the difficulty of data
measurement in cryogenic flows, experimental studies on cryo-
genic cavitation are comparatively difficult to conduct and there
is still limited research available when compared with typical
cavitation studies of water. With the development of experi-
mental techniques, such as the setup of a cryogenic experimen-
tal platform and the utilization of advanced measurement
techniques for the cryogenic cavitation flow field, the under-
standing of the characteristics and mechanism of cryogenic cav-
itation has been greatly promoted. Owing to the vacuum
sealing and thermal insulation technique, cavitating flow visual-
izations, as recorded photographically by a high-speed camera,
have been one of the most important research methods for
cryogenic cavitation study, particularly in the identification of
cavitation morphology and determination of cavity dynamics.
The temperature depression, which is the defining feature of
cavitation in cryogenic fluids, has been specially measured. In
addition, the pressure and velocity of the cavity, which are of
great significance for physically deepening the understanding of
cavitation, have also been generally measured to obtain the
unsteady characteristics of cryogenic cavitation. Because of the
technological limitations, cryogenic cavitating flow experiments
FIG. 24. Contours of the pressure field at distinct instants.161 Reproduced with per- in the inducer or turbopump are extremely difficult to perform,
mission from “Investigation of unsteady cryogenic cavitating flow and induced noise and very few experiments on cryogenic cavitation use LH2 as a
around a three-dimensional hydrofoil,” Phys. Fluids 34, 042120 (2022). Copyright working fluid.
2022 AIP Publishing. (2) Numerical simulation has become an indispensable tool for
22 November 2024 04:35:04

resolving the complex multi-scale multiphase problems of cryo-
studied in which geometry, and what is the focus of the study. genic cavitation. To accurately predict the behavior of cryogenic
Combined with the above detailed introduction and analysis, the cavitation, the cavitation model and turbulence model have
research progress and existing problems of cryogenic cavitation can been modified and calibrated to agree with the reported experi-
be summarized in Fig. 25. mental data. With regard to thermal effects modeling, the
FIG. 25. Summary of cryogenic cavitation experiment and simulation studies.
Phys. Fluids 34, 101303 (2022); doi: 10.1063/5.0102876 34, 101303-35

energy equation considering the cavitation phase-change effects which can capture the formation, development, and breakup of sheet
should be included and the variations of physical properties cavitation and the shedding process of cloud cavitation, can be intro-
and vapor pressure as a result of temperature drop need to be duced to better characterize the microscopic behavior of the cryogenic
considered. The transport-equation-based cavitation model has cavitation. On the other hand, an effective numerical framework of
been typically corrected to include the influences of thermal cryogenic cavitation fluid–structure interaction needs urgent develop-
effect on the bubble growth and collapse. On the other hand, ment to investigate the characteristics of unsteady cavitating flow-
the turbulence model also plays an important role in the induced vibration in the turbopump.
numerical modeling of cryogenic cavitation flow. With the con-
tinuous improvement of computer performance, the advantages ACKNOWLEDGMENTS
of LES for cavitation mechanism investigations have gradually
become prominent, but the influence of grids on the calculation This work was supported by the Nature and Science Foundation
results needs to be carefully checked. of China (No. 51976177) and the National Key Research and
(3) Great efforts have been made to study the thermal effects of Development Program of China (No. 2021YFB4000700).
cryogenic cavitation. The distinctive cavitation morphology for
cryogenic cavitation under the dominance of thermal effects AUTHOR DECLARATIONS
has been well presented. The temperature drop characteristics Conflict of Interest
as well as the influences of thermal effects on the cavitation
dynamics, inducer cavitation, and pressure fluctuation have The authors have no conflicts to disclose.
been made great progress.
(4) Investigations into the shedding mechanisms of cryogenic cavi- Author Contributions
tation remain at an early stage and a late start compared to cav- Aibo Wei: Conceptualization (equal); Data curation (equal);
itation in water. Both the reentrant mechanism and shock wave Investigation (lead); Software (lead); Validation (equal); Visualization
mechanism have been observed in the cryogenic cavitation, but (lead); Writing – original draft (lead). Lianyan Yu: Data curation
there still lacks enough experimental measurements and the (equal); Formal analysis (equal); Validation (equal); Visualization
corresponding analysis to enhance the understanding of the (supporting); Writing – review and editing (equal). Limin Qiu:
behaviors of reentrant jet and shock wave. In addition, a special Formal analysis (equal); Methodology (equal); Project administration
22 November 2024 04:35:04

shedding mechanism of the interaction between cavitation and (equal); Supervision (supporting). Xiaobin Zhang: Conceptualization
vortex accompanied by thermal effects has been proposed. The (equal); Funding acquisition (lead); Methodology (equal); Project
cavitation–vortex interactions of cryogenic cavitation illustrate administration (equal); Supervision (lead); Writing – review and edit-
a quite different feature due to the liquid entrainment in the ing (equal).
cavity, which has become the focus of cryogenic cavitating field
analysis.
(5) Cavitation has presented a persistent challenge to the design of DATA AVAILABILITY
inducers and turbopumps because of the important impact of The data that support the findings of this study are available
unsteady cavitating flow-induced vibration and noise. However, from the corresponding author upon reasonable request.
it is of great difficulty to conduct related experiments, and few
numerical studies on this topic have been reported due to the REFERENCES
highly complex fluid–structure interactions associated with 1
L. Xiang, H. Chen, Y. Tan, K. Xu, and J. Liu, “Study of thermodynamic cavita-
strong phase transition and a large span of space scales and tion effects in an inducer,” J. Propul. Power 36, 312 (2020).
timescales. 2
K. Mani, A. Cervone, and J.-P. Hickey, “Turbulence modelling and cavitation
dynamics in cryogenic turbopumps,” in Proceedings of the 5th Space
Finally, future research expectations on the cryogenic cavitation Propulsion (2016).
are proposed as follows: 3
R. Xue, L. Chen, X. Zhong, X. Liu, S. Chen, and Y. Hou, “Unsteady cavitation
In terms of experimental studies, continuing desire for dynamic of liquid nitrogen flow in spray nozzles under fluctuating conditions,”
measurements of pressure, velocity, and temperature variations in the Cryogenics 97, 144 (2019).
cavitation zone needs to be met to further reveal the unsteady charac-
4
N. Huang, Z. Li, and B. Zhu, “Cavitating flow suppression in the draft tube of
teristics of cryogenic cavitation. The measurement of vapor distribu- a cryogenic turbine expander through runner optimization,” Processes 8, 270
(2020).
tion is of great significance to the analysis of cavity structure, and the 5
A.-B. Wei, R. Gao, W. Zhang, S.-H. Wang, R. Zhou, and X.-B. Zhang,
vapor distribution measurement method such as x-ray will greatly “Computational fluid dynamics analysis on flow-induced vibration of a cryo-
facilitate the identification and analysis of cryogenic cavity structure. genic poppet valve in consideration of cavitation effect,” J. Zhejiang Univ. Sci.
In addition, the PIV/PLIF can simultaneously obtain flow field infor- A 23, 83 (2022).
mation inside and outside the cavity, which should be emphasized and 6
Y. Utturkar, J. Wu, G. Wang, and W. Shyy, “Recent progress in modeling of
developed in future experimental research. cryogenic cavitation for liquid rocket propulsion,” Prog. Aerosp. Sci. 41, 558
While for numerical studies, the currently widely used cavitation (2005).
7
W. Li, S. Li, L. Ji, X. Zhao, W. Shi, R. K. Agarwal, M. Awais, and Y. Yang, “A
models for cryogenic cavitation modeling can only solve the vapor vol- study on the cavitation and pressure pulsation characteristics in the impeller
ume fraction distribution at the macroscopic level, the microscopic of an LNG submerged pump,” Machines 10, 14 (2021).
mechanism of cavitation evolution cannot be depicted. A multi-scale 8
D.-J. Kim, H. J. Sung, C.-H. Choi, and J.-S. Kim, “Cavitation instabilities of
cavitation model based on the Euler–Lagrange coupling method, an inducer in a cryogenic pump,” Acta Astronaut. 132, 19 (2017).
Phys. Fluids 34, 101303 (2022); doi: 10.1063/5.0102876 34, 101303-36

9
M. C. Vilela Salgado, C. M. N. Belderrain, and T. C. Devezas, “Space propul- 36
T. Toshimasa, H. Eizo, T. Kazuhiro, and M. Kazuo, “Study of laser-induced
sion: A survey study about current and future technologies,” J. Aerosp. cavitation bubble in liquid nitrogen,” Proc. SPIE 5777, XV International
Technol. Manage. 10, e1118 (2018). Symposium on Gas Flow, Chemical Lasers, and High-Power Lasers (2005).
10
K. J. Petersen and J. R. Brinkerhoff, “On the lattice Boltzmann method and its 37
M. Petkovšek and M. Dular, “IR measurements of the thermodynamic effects
application to turbulent, multiphase flows of various fluids including cryo- in cavitating flow,” Int. J. Heat Fluid Flow 44, 756 (2013).
gens: A review,” Phys. Fluids 33, 041302 (2021). 38
L. R. Sarosdy and A. J. Acosta, “Note on observations of cavitation in different
11
M. G. D. Giorgi, D. Bello, and A. Ficarella, “Analysis of thermal effects in a fluids,” J. Basic Eng. 83, 399 (1961).
cavitating orifice using Rayleigh equation and experiments,” J. Eng. Gas 39
Y. Ito, K. Seto, and T. Nagasaki, “Periodical shedding of cloud cavitation from
Turbines Power 132, 092901 (2010). a single hydrofoil in high-speed cryogenic channel flow,” J. Therm. Sci. 18, 58
12
J.-P. Franc and C. Pellone, “Analysis of thermal effects in a cavitating inducer (2009).
using Rayleigh equation,” J. Fluids Eng. 129, 974 (2007). 40
X. Gao, H. Chen, J. Zhu, S. Wang, and X. Zhang, “Visual experimental study
13
M. Petkovsek and M. Dular, “Cavitation dynamics in water at elevated tem- on liquid-nitrogen cavitating flow on NACA 66 hydrofoil,” J. Propul. Power
peratures and in liquid nitrogen at an ultrasonic horn tip,” Ultrason. 36, 88 (2020).
Sonochem. 58, 104652 (2019). 41
T. Chen, H. Chen, W. Liang, B. Huang, and L. Xiang, “Experimental investiga-
14
H. A. Stahl and A. J. Stepanoff, “Thermodynamic aspects of cavitation in cen- tion of liquid nitrogen cavitating flows in converging-diverging nozzle with
trifugal pumps,” J. Fluids Eng. 78, 1691 (1956). special emphasis on thermal transition,” Int. J. Heat Mass Transfer 132, 618
15
M. L. Billet, J. W. Holl, and D. S. Weir, “Correlations of thermodynamic (2019).
effects for developed cavitation,” J. Fluids Eng. 103, 534 (1981). 42
T. Chen, H. Chen, W. Liu, B. Huang, and G. Wang, “Unsteady characteristics
16
J. W. Holl, M. L. Billet, and D. S. Weir, “Thermodynamic effects on developed of liquid nitrogen cavitating flows in different thermal cavitation mode,”
cavitation,” J. Fluids Eng. 97, 507 (1975). Appl. Therm. Eng. 156, 63 (2019).
17
T. F. Gelder, R. D. Moore, and R. S. Ruggeri, “Cavitation similarity consider- 43
W. Liang, T. Chen, G. Wang, and B. Huang, “Experimental investigations on
ations based on measured pressure and temperature depressions in cavitated transient dynamics of cryogenic cavitating flows under different free-stream
regions of freon 114,” Report No. NASA TN D-3509, 1966. conditions,” Int. J. Heat Mass Transfer 178, 121537 (2021).
18
R. D. Moore and R. S. Ruggeri, “Prediction of thermodynamic effects of devel- 44
M. Murakami and K. Harada, “Thermo-fluid dynamic experiment of He II
oped cavitation based on liquid-nitrogen and freon 114 data in scaled ventu- cavitating flow,” in CAV2009 7th International Symposium on Cavitation,
ris,” Report No. NASA TN D-4899, 1962. Ann Arbor, MI, 2009 (Curran, 2011).
19
J. Hord, “Cavitation in liquid cryogens. II. Hydrofoil,” Report No. NASA-CR- 45
M. Oike, M. Kikuchi, S. Takada, T. Sudo, and T. Takano, “Effects of the recess
2156, 1973. geometry on flow characteristics of cryogenic hybrid journal bearings for
20
H. Rouse and J. S. McNown, Cavitation and Pressure Distribution: Head rocket turbopumps,” Tribol. Online 13, 140 (2018).
Forms at Zero Angle of Yaw (State University of Iowa, 1948). 46
C. L. Ball, P. R. Meng, and L. Reed, “Cavitation performance of 84 helical
22 November 2024 04:35:04

21
D. H. Fruman, I. Benmansour, and R. Sery, “Estimation of the thermal effects pump inducer operated in 37 and 42 R liquid hydrogen,” Report No.
on cavitation of cryogenic liquids,” in Cavitation and Multiphase Flow Forum, NASA-TM-X-1360, 1967.
47
ASME FED, Toronto, Canada, 4–7 June 1990 (ASME, 1991). Y. Yoshida, Y. Sasao, M. Watanabe, T. Hashimoto, Y. Iga, and T. Ikohagi,
22
D. H. Fruman, J.-L. Reboud, and B. Stutz, “Estimation of thermal effects in cav- “Thermodynamic effect on rotating cavitation in an inducer,” J. Fluids Eng.
itation of thermosensible liquids,” Int. J. Heat Mass Transfer 42, 3195 (1999). 131, 091302 (2009).
23
E. R. G. Eckert, Jr. and R. M. Drake, Analysis of Heat and Mass Transfer 48
Y. Ito, A. Tsunoda, Y. Kurishita, S. Kitano, and T. Nagasaki, “Experimental
(Hemisphere Publishing Corporation, 1987). visualization of cryogenic backflow vortex cavitation with thermodynamic
24
Y. Utturkar, “Computational modeling of thermodynamic effects in cryogenic effects,” J. Propul. Power 32, 71 (2016).
cavitation,” Ph.D. thesis (University of Florida, 2005). 49
Y. Ito, “The world’s first test facility that enables the experimental visualiza-
25
C. E. Brennen, Fundamentals of Multiphase Flow (Cambridge University tion of cavitation on a rotating inducer in both cryogenic and ordinary flu-
Press, Cambridge, UK, 2005). ids,” J. Fluids Eng. 143, 121105 (2021).
26
Z. Zuo, H. Zhang, Z. Ren, H. Chen, and S. Liu, “Thermodynamic effects at 50
J. Ishimoto, M. Onishi, and K. Kamijo, “Numerical and experimental study
Venturi cavitation in different liquids,” Phys. Fluids 34, 083310 (2022). on the cavitating flow characteristics of pressurized liquid nitrogen in a hori-
27
M. S. Plesset and A. Prosperetti, “Bubble dynamics and cavitation,” Annu. zontal rectangular nozzle,” J. Pressure Vessel Technol. 127, 515 (2004).
Rev. Fluid Mech. 9, 145 (1977). 51
K. Ohira, T. Nakayama, and T. Nagai, “Cavitation flow instability of sub-
28
J.-P. Franc, C. Rebattet, and A. Coulon, “An experimental investigation of cooled liquid nitrogen in converging–diverging nozzles,” Cryogenics 52, 35
thermal effects in a cavitating inducer,” J. Fluids Eng. 126, 716 (2004). (2012).
29
D. A. Ehrlich and J. W. Murdock, “A dimensionless scaling parameter for 52
J. Zhu, H. Xie, K. Feng, X. Zhang, and M. Si, “Unsteady cavitation characteris-
thermal effects on cavitation in turbopump inducers,” J. Fluids Eng. 137, tics of liquid nitrogen flows through venturi tube,” Int. J. Heat Mass Transfer
042203 (2015). 112, 544 (2017).
30
T. Chen, B. Huang, G. Wang, and X. Zhao, “Numerical study of cavitating 53
J. Zhu, S. Wang, and X. Zhang, “Influences of thermal effects on cavitation
flows in a wide range of water temperatures with special emphasis on two dynamics in liquid nitrogen through venturi tube,” Phys. Fluids 32, 012105
typical cavitation dynamics,” Int. J. Heat Mass Transfer 101, 886 (2016). (2020).
31
J. Hord, “Cavitation in liquid cryogens. III. Ogives,” Report No. NASA-CR- 54
T. Chen, Z. Mu, B. Huang, M. Zhang, and G. Wang, “Dynamic instability
2242, 1973. analysis of cavitating flow with liquid nitrogen in a converging–diverging
32
J. Hord, “Cavitation in liquid cryogens. IV. Combined correlations for ven- nozzle,” Appl. Therm. Eng. 192, 116870 (2021).
turi, hydrofoil, ogives, and pumps,” Report No. NASA-CR-2448, 1974. 55
C. Esposito, “Study of cryogenic transient flows. The impact of the fluid ther-
33
H. Sato, X. W. Sun, M. Qdagawa, K. Maeno, and H. Honma, “An investiga- mosensitivity on cavitation,” Ph.D. thesis (Catholic University of Leuven,
tion on the behavior of laser induced bubble in cryogenic liquid nitrogen,” 2020).
J. Fluids Eng. 118, 850 (1996). 56
C. Esposito, L. Peveroni, J. B. Gouriet, J. Steelant, and M. R. Vetrano, “On the
34
S. Watanabe, T. Sirato, M. Ota, and K. Maeno, “Bubble dynamics induced by influence of thermal phenomena during cavitation through an orifice,” Int. J.
YAG laser focusing in liquid nitrogen and cryogenic laser processing for par- Heat Mass Transfer 164, 120481 (2021).
ticles,” Proc. SPIE 6346, 63461R (2007). 57
N. Fdida, B. Patte-Rouland, J. B. Blaisot, E. Rouland, and V. Begin, “Velocity
35
K. Maeno, H. Sato, and S. Endo, “Some laser applications to the fluid materi- measurements of bubbles in liquid nitrogen,” in Proceedings of the 23rd
als: Laser-induced cavitation in cryogenic liquids and gas decomposition by Annual Conference on Liquid Atomization and Spray Systems (2010).
laser,” in Tenth International School on Quantum Electronics: Laser Physics 58
M. Dular and M. Petkovšek, “Cavitation erosion in liquid nitrogen,” Wear
and Applications (Society of Photo Optical, 1999). 400–401, 111 (2018).
Phys. Fluids 34, 101303 (2022); doi: 10.1063/5.0102876 34, 101303-37

59 83
Y. Yoshida, M. Eguchi, T. Motomura, M. Uchiumi, H. Kure, and Y. Maruta, R. B. Medvitz, R. F. Kunz, D. A. Boger, J. W. Lindau, A. M. Yocum, and L. L.
“Rotordynamic forces acting on three-bladed inducer under supersynchro- Pauley, “Performance analysis of cavitating flow in centrifugal pumps using
nous/synchronous rotating cavitation,” J. Fluids Eng. 132, 061105 (2010). multiphase CFD,” J. Fluids Eng. 124, 377 (2003).
60
A. Wei, L. Yu, R. Gao, W. Zhang, and X. Zhang, “Unsteady cloud cavitation 84
J. Zhu, Y. Chen, D. Zhao, and, X. Zhang, “Extension of the Schnerr-Sauer
mechanisms of liquid nitrogen in convergent–divergent nozzle,” Phys. Fluids model for cryogenic cavitation,” European J. Mechanics B/Fluids 52, 1 (2015).
33, 092116 (2021). 85
J. Zhu, S. Wang, L. Qiu, X. Zhi, and X. Zhang, “Frequency characteristics of
61
C. Esposito, L. Peveroni, J. B. Gouriet, J. Steelant, and M. R. Vetrano, liquid hydrogen cavitating flow over a NACA0015 hydrofoil,” Cryogenics 90,
“Hydraulic similitude assessment for cryogenic cavitation in propellant lines: 7 (2018).
The case of thick orifice,” Cryogenics 121, 103407 (2022). 86
S. Wang, J. Zhu, H. Xie, F. Zhang, and X. Zhang, “Studies on thermal effects
62
J. Pinho, L. Peveroni, M. R. Vetrano, J.-M. Buchlin, J. Steelant, and M. of cavitation in LN2 flow over a twisted hydrofoil based on large eddy simula-
Strengnart, “Experimental and numerical study of a cryogenic valve using liq- tion,” Cryogenics 97, 40 (2019).
uid nitrogen and water,” Aerosp. Sci. Technol. 93, 105331 (2019). 87
T. Sun, X. Ma, Y. Wei, and C. Wang, “Computational modeling of cavitating
63
J. L. Mena, M. A. Ingle, V. Shirsat, and A. Choudhuri, “An investigation of a flows in liquid nitrogen by an extended transport-based cavitation model,”
cavitating venturi flow control feature in a cryogenic propellant delivery sys- Sci. China Technol. Sci. 59, 337 (2016).
tem,” Flow Meas. Instrum. 41, 97 (2015). 88
X. Li, T. Shen, P. Li, X. Guo, and Z. Zhu, “Extended compressible thermal
64
M. Murakami and K. Harada, “Experimental study of thermo-fluid dynamic cavitation model for the numerical simulation of cryogenic cavitating flow,”
effect in He II cavitating flow,” Cryogenics 52, 620 (2012). Int. J. Hydrogen Energy 45, 10104 (2020).
65
Y. Ito, T. Nagayama, and T. Nagasaki, “Cavitation patterns on a plano convex 89
X. Zhang, J. Zhu, L. Qiu, and X. Zhang, “Calculation and verification of
hydrofoil in a highspeed cryogenic cavitation tunnel,” in CAV2009 7th dynamical cavitation model for quasi-steady cavitating flow,” Int. J. Heat
International Symposium on Cavitation, Ann Arbor, MI 2009 (Curran, 2011). Mass Transfer 86, 294 (2015).
66
K. Niiyama, S. Hasegawa, S. Tsuda, Y. Yoshida, T. Tamura, and M. Oike, 90
S. Shi and G. Wang, “Numerical calculation of thermal effect on cavitation in
“Thermodynamic effects on cryogenic cavitating flow in an orifice,” in cryogenic fluids,” Chin. J. Mech. Eng. 25, 1176 (2012).
CAV2009 7th International Symposium on Cavitation, Ann Arbor, MI 2009 91
S. Shi and G. Wang, “Numerical analysis of factors influencing thermal effects
(Curran, 2011). generated by cavitation flow of cryogenic fluids,” Mod. Phys. Lett. B 34,
67
E. Goncalves and R. F. Patella, “Numerical simulation of cavitating flows with 2050184 (2020).
homogeneous models,” Comput. Fluids 38, 1682 (2009). 92
C.-C. Tseng, Y. Wei, G. Wang, and W. Shyy, “Modeling of turbulent, isother-
68
E. Goncalves and R. F. Patella, “Numerical study of cavitating flows with thermal and cryogenic cavitation under attached conditions,” Acta Mech. Sin. 26,
modynamic effect,” Comput. Fluids 39, 99 (2010). 325 (2010).
69
P. Cooper, “Analysis of single-and two-phase flows in turbopump inducers,” 93
J. Wu, G. Wang, and W. Shyy, “Time-dependent turbulent cavitating flow
J. Eng. Power 89, 577 (1967). computations with interfacial transport and filter-based models,” Int. J.
22 November 2024 04:35:04

70
Y. ITO, K. Sawasaki, N. Tani, T. Nagasaki, and T. Nagashima, “A blowdown Numer. Methods Fluids 49, 739 (2005).
cryogenic cavitation tunnel and CFD treatment for flow visualization around 94
Y. Liu, X. Li, Z. Lin, L. Li, and Z. Zhu, “Numerical analysis of thermo-
a foil,” J. Therm. Sci. 14, 346 (2005). sensitive cavitating flows with special emphasises on flow separation and ens-
71
C. Merkle, J. Feng, and P. Buelow, “Computational modeling of the dynamics trophy conversion,” Int. Commun. Heat Mass Transfer 125, 105336 (2021).
of sheet cavitaiton,” in Proceedings of the 3rd International Symposium on 95
S. Sun, J. Sun, W. Sun, and P. Song, “Enhancing cryogenic cavitation predic-
Cavitation, Grenoble, France, 7–10 April 1998 (1998). tion through incorporating modified cavitation and turbulence models,”
72
R. F. Kunz, D. A. Boger, D. R. Stinebring et al., “A preconditioned J. Fluids Eng. 143, 061404 (2021).
Navier–Stokes method for two-phase flows with application to cavitation pre- 96
O. Coutier-Delgosha, R. Fortes-Patella, and J.-L. Reboud, “Evaluation of the
diction,” Comput. Fluids 29, 849 (2000). turbulence model influence on the numerical simulations of unsteady cavita-
73
G. H. Schnerr and J. Sauer, “Physical and numerical modeling of unsteady tion,” J. Fluids Eng. 125, 38 (2003).
cavitation dynamics,” in Fourth International Conference on Multiphase Flow 97
X. Long, Q. Liu, B. Ji, and Y. Lu, “Numerical investigation of two typical cavi-
(Elsevier, 2001). tation shedding dynamics flow in liquid hydrogen with thermodynamic
74
A. K. Singhal and H. Li et al., “Mathematical basis and validation of the full effects,” Int. J. Heat Mass Transfer 109, 879 (2017).
cavitation model,” J. Fluids Eng. 124, 617 (2002). 98
B. Huang, Q. Wu, and G. Wang, “Numerical investigation of cavitating flow
75
I. Senocak and W. Shyy, “Interfacial dynamics-based modelling of turbulent in liquid hydrogen,” Int. J. Hydrogen Energy 39, 1698 (2014).
cavitating flows. II. Time-dependent computations,” Int. J. Numer. Methods 99
W. Liang, T. Chen, G. Wang, and B. Huang, “Investigation of unsteady liquid
Fluids 44, 997 (2004). nitrogen cavitating flows with special emphasis on the vortex structures using
76
I. Senocak and W. Shyy, “A pressure-based method for turbulent cavitating mode decomposition methods,” Int. J. Heat Mass Transfer 157, 119880
flow computations,” J. Comput. Phys. 176, 363 (2002). (2020).
77
X. Zhang, Z. Wu, S. Xiang, and L. Qiu, “Modeling cavitation flow of cryogenic 100
J. Zhu, D. Zhao, L. Xu, and X. Zhang, “Interactions of vortices, thermal effects
fluids with thermodynamic phase-change theory,” Chin. Sci. Bull. 58, 567 and cavitation in liquid hydrogen cavitating flows,” Int. J. Hydrogen Energy
(2013). 41, 614 (2016).
78
P. J. Zwart, A. G. Gerber, and T. Belamri, “A two-phase flow model for pre- 101
J. Smagorinsky, “General circulation experiments with the primitive equations.
dicting cavitation dynamics,” in 5th International Conference on Multiphase I. The basic experiment,” Mon. Weather Rev. 91, 99 (1963).
Flow (ICMF, 2004). 102
F. Nicoud and F. Ducros, “Subgrid-scale stress modelling based on the square
79
T. Goel, S. Thakur, R. T. Haftka, W. Shyy, and J. Zhao, “Surrogate model- of the velocity gradient tensor,” Flow Turbul. Combust. 62, 183 (1999).
based strategy for cryogenic cavitation model validation and sensitivity evalu- 103
W.-W. Kim and S. Menon, “Application of the localized dynamic subgrid-
ation,” Int. J. Numer. Methods Fluids 58, 969 (2008). scale model to turbulent wall-bounded flows,” AIAA Paper No. 1997-210,
80
M. G. De Giorgi, A. Ficarella, and D. Fontanarosa, “Impact of population bal- 1997.
ance modeling on the prediction of cryogenic cavitation in aerospace propul- 104
M. Germano, U. Piomelli, P. Moin, and W. H. Cabot, “A dynamic subgrid-
sion systems,” AIAA Paper No. 2018-4859, 2018. scale eddy viscosity model,” Phys. Fluids A: Fluid Dyn. 3, 1760 (1991).
81
M. G. D. Giorgi, A. Ficarella, and D. Laforgia, “Modeling nucleation phenom- 105
D. K. Lilly, “A proposed modification of the Germano subgrid-scale closure
ena in cavitating flow,” AIAA Paper No. 2007-4459, 2007. method,” Phys. Fluids A: Fluid Dyn. 4, 633 (1992).
82
M. G. De Giorgi, A. Ficarella, and D. Fontanarosa, “Numerical investigation 106
S.-E. Kim, “Large eddy simulation using unstructured meshes and dynamic
of nonisothermal cavitating flows on hydrofoils by means of an extended subgrid-scale turbulence models,” AIAA Paper No. 2004-2548, 2004.
Schnerr-Sauer model coupled with a nucleation model,” J. Eng. Gas Turbines 107
A. Hosangadi and V. Ahuja, “A new unsteady model for dense cloud cavita-
Power 142, 041003 (2020). tion in cryogenic fluids,” AIAA Paper No. 2005-5347, 2005.
Phys. Fluids 34, 101303 (2022); doi: 10.1063/5.0102876 34, 101303-38

108
C.-C. Tseng and W. Shyy, “Modeling for isothermal and cryogenic cavitation,” 132
S. Venkateswaran, J. W. Lindau, R. F. Kunz, and C. L. Merkle, “Computation
Int. J. Heat Mass Transfer 53, 513 (2009). of multiphase mixture flows with compressibility effects,” J. Comput. Phys.
109
H. Kim and C. Kim, “A physics-based cavitation model ranging from inertial 180, 54 (2002).
to thermal regimes,” Int. J. Heat Mass Transfer 181, 121991 (2021). 133
A. D. Le, J. Okajima, and Y. Iga, “Numerical simulation study of cavitation in
110
Y. Liu, X. Li, M. Ge, L. Li, and Z. Zhu, “Numerical investigation of transient liquefied hydrogen,” Cryogenics 101, 29 (2019).
134
liquid nitrogen cavitating flows with special emphasis on force evolution and Y. Yoshida, K. Kikuta, S. Hasegawa, M. Shimagaki, and T. Tokumasu,
entropy features,” Cryogenics 113, 103225 (2021). “Thermodynamic effect on a cavitating inducer in liquid nitrogen,” J. Fluids
111
S. Shi, G. Wang, and C. Hu, “A Rayleigh-Plesset based transport model for Eng. 129, 273 (2007).
135
cryogenic fluid cavitating flow computations,” Sci. China Phys. Mech. Astron. Y. Yoshida, Y. Sasao, K. Okita, S. Hasegawa, M. Shimagaki, and T. Ikohagi,
57, 764 (2014). “Influence of thermodynamic effect on synchronous rotating cavitation,” J. Fluids
112
H. Chang, X. Xie, Y. Zheng, and S. Shu, “Numerical study on the cavitating Eng. 129, 871 (2007).
flow in liquid hydrogen through elbow pipes with a simplified cavitation 136
X.-Y. Duan, X. Song, K.-X. Chu, C.-Y. Zhu, L. Gong, and M. Xu, “Numerical inves-
model,” Int. J. Hydrogen Energy 42, 18325 (2017). tigation of the cryogenic cavitating flow characteristics in bi-directional cone flow-
113
D. Li, B. Miao, Y. Li, R. Gong, and H. Wang, “Numerical study of the hydrofoil meters: A systematic parameter study,” Heat Transfer Eng. 42, 1348 (2021).
cavitation flow with thermodynamic effects,” Renewable Energy 169, 894 137
T. Chen, H. Chen, B. Huang, W. Liang, L. Xiang, and G. Wang, “Thermal
(2021). transition and its evaluation of liquid hydrogen cavitating flow in a wide range
114
S.-I. Tsuda, N. Tani, and N. Yamanishi, “Development and validation of a of free-stream conditions,” Int. J. Heat Mass Transfer 127, 1277 (2018).
reduced critical radius model for cryogenic cavitation,” J. Fluids Eng. 134, 138
W. Liang, T. Chen, B. Huang, and G. Wang, “Thermodynamic analysis of
051301 (2012). unsteady cavitation dynamics in liquid hydrogen,” Int. J. Heat Mass Transfer
115
Y. Iga, M. Nohmi, A. Goto, B. R. Shin, and T. Ikohagi, “Numerical study of 142, 118470 (2019).
sheet cavitation breakoff phenomenon on a cascade hydrofoil,” J. Fluids Eng. 139
R. S. Ruggeri and R. D. Moore, “Method for prediction of pump cavitation
125, 643 (2003). performance for various liquids,” Report No. NASA-TN-D-5292, 1969.
116
Y. Utturkar, S. Thakur, and W. Shyy, “Computational modeling of thermody- 140
G. Kovich, “Comparison of predicted and experimental cavitation perfor-
namic effects in cryogenic cavitation,” AIAA Paper No. 2005-1286, 2005. mance of 84 helical inducer in water and hydrogen,” Report No. NASA-TN-
117
V. Ahuja, A. Hosangadi, and J. Shipman, “Computational analyses of cavitat- D-7016, 1970.
141
ing control elements in cryogenic environments,” in Proceedings of the ASME K. Kikuta, Y. Yoshida, T. Hashimoto, H. Nanri, T. Mizuno, and N. Shimiya,
2004 Heat Transfer/Fluids Engineering Summer Conference (ASME, Charlotte, “Influence of rotational speed on thermodynamic effect in a cavitating
NC, 2004). inducer,” in Proceedings of the ASME 2009 Fluids Engineering Division
118
A. D. Le, O. Junosuke, and I. Yuka, “Modification of energy equation for Summer Meeting (ASME, Colorado, 2009).
homogeneous cavitation simulation with thermodynamic effect,” J. Fluids Eng. 142
A. Hosangadi, V. Ahuja, and R. J. Ungewitter, “Simulations of cavitating cryo-
22 November 2024 04:35:04

141, 081102 (2019). genic inducers,” AIAA Paper No. 2004-4023, 2004.
119
S. Zhang, X. Li, and Z. Zhu, “Numerical simulation of cryogenic cavitating 143
E. Goncalves, R. F. Patella, J. Rolland, B. Pouffary, and G. Challier,
flow by an extended transport based cavitation model with thermal effects,” “Thermodynamic effect on a cavitating inducer in liquid hydrogen,” J. Fluids
Cryogenics 92, 98 (2018). Eng. 132, 111305 (2010).
120
T. Chen, G. Wang, B. Huang, and K. Wang, “Numerical study of thermody- 144
R. T. Knapp, “Cavitation and nuclei,” Trans. Am. Soc. Mech. Eng. 80, 1315
namic effects on liquid nitrogen cavitating flows,” Cryogenics 70, 21 (2015). (1958).
121
S. Rahbarimanesh, J. Brinkerhoff, and J. Huang, “Development and validation 145
H. Ganesh, S. A. M€akiharju, and S. L. Ceccio, “Bubbly shock propagation as a
of a homogeneous flow model for simulating cavitation in cryogenic fluids,” mechanism for sheet-to-cloud transition of partial cavities,” J. Fluid Mech.
Appl. Math. Modell. 56, 584 (2018). 802, 37 (2016).
122
T. Chen, B. Huang, G. Wang, and K. Wang, “Effects of fluid thermophysical 146
M. Petkovšek, M. Hocevar, and M. Dular, “Visualization and measurements
properties on cavitating flows,” J. Mech. Sci. Technol. 29, 4239 (2015). of shock waves in cavitating flow,” Exp. Therm. Fluid Sci. 119, 110215 (2020).
123
S. Rahbarimanesh, “Numerical investigation of cryogenic cavitation in lique- 147
T. Sun, Y. Wei, L. Zou, Y. Jiang, C. Xu, and Z. Zong, “Numerical investigation
fied natural gas (LNG) flows using a homogeneous flow model,” Ph.D. thesis on the unsteady cavitation shedding dynamics over a hydrofoil in thermo-
(University of British Columbia, 2020). sensitive fluid,” Int. J. Multiphase Flow 111, 82 (2019).
124
C. Shao, Z. Zhang, and J. Zhou, “Study of the flow in a cryogenic pump under 148
S. Zhang, X. Li, B. Hu, Y. Liu, and Z. Zhu, “Numerical investigation of attached
different cavitation inducements by considering the thermodynamic effect,” cavitating flow in thermo-sensitive fluid with special emphasis on thermal effect
Int. J. Numer. Methods Heat and Fluid Flow 30, 4307 (2019). and shedding dynamics,” Int. J. Hydrogen Energy 44, 3170 (2019).
125
D. Li, Z. Ren, Y. Li, R. Gong, and H. Wang, “Thermodynamic effects on the 149
B. Ji, X. Luo, R. E. A. Arndt, and Y. Wu, “Numerical simulation of three
cavitation flow of a liquid oxygen turbopump,” Cryogenics 116, 103302 dimensional cavitation shedding dynamics with special emphasis on cavita-
(2021). tion–vortex interaction,” Ocean Eng. 87, 64 (2014).
126
X. Zhang, L. Qiu, H. Qi, X. Zhang, and Z. Gan, “Modeling liquid hydrogen 150
C.-Z. Han, S. Xu, H.-Y. Cheng, B. Ji, and Z.-Y. Zhang, “LES method of the tip
cavitating flow with the full cavitation model,” Int. J. Hydrogen Energy 33, clearance vortex cavitation in a propelling pump with special emphasis on the
7197 (2008). cavitation-vortex interaction,” J. Hydrodyn. 32, 1212 (2020).
127
X. B. Zhang, L. M. Qiu, Y. Gao, and X. J. Zhang, “Computational fluid 151
Y. Long, X. Long, and B. Ji, “LES investigation of cavitating flows around a
dynamic study on cavitation in liquid nitrogen,” Cryogenics 48, 432 (2008). sphere with special emphasis on the cavitation–vortex interactions,” Acta
128
X. Cao, X. Zhang, L. Qiu, and Z. Gan, “Validation of full cavitation model in Mech. Sin. 36, 1238 (2020).
cryogenic fluids,” Sci. Bull. 54, 1633 (2009). 152
Y. Yamamoto, S. Nakajima, S. Watanabe, M. Ota, and K. Maeno, “Single bub-
129
B. Xu, J. Feng, F. Wan, D. Zhang, X. Shen, and W. Zhang, “Numerical investi- ble behavior induced by Nd:YAG laser focusing near solid wall in liquid nitro-
gation of modified cavitation model with thermodynamic effect in water and gen,” Trans. Jpn. Soc. Aeronaut. Space Sci. Technol. Jpn. 7, Pe_65 (2009).
liquid nitrogen,” Cryogenics 106, 103049 (2020). 153
A. Mishra, J. Mondal, A. Roy, R. Lakkaraju, and P. Ghosh, Effect of Curved
130
B. Xu, J. Feng, X. Shen, D. Zhang, and W. Zhang, “Numerical investigation of Rigid Surface on the Collapsing Cavitating Bubble in Cryogenic Environment
cavitation suppression in an inducer for water and liquid nitrogen with (IOP Publishing, 2020).
emphasis on thermodynamic effect,” J. Braz. Soc. Mech. Sci. Eng. 43, 212 154
Y. Yoshida, Y. Kazami, K. Nagaura, M. Shimagaki, Y. Iga, and T. Ikohagi,
(2021). “Interaction between uneven cavity length and shaft vibration at the inception
131
E. Rapposelli and L. d’Agostino, “A barotropic cavitation model with thermo- of synchronous rotating cavitation,” Int. J. Rotating Mach. 2008, 1.
155
dynamic effects,” in Proceedings of the Fifth International Symposium on T. Shimura, M. Shimagaki, Y. Watanabe, M. Watanabe, S. Hasegawa, and S.
Cavitation (Osaka, Japan, 2003). Takata, “Cavitation induced vibration of LE 7A oxygen turpopump,”
Phys. Fluids 34, 101303 (2022); doi: 10.1063/5.0102876 34, 101303-39

CAV2003, 5th International Symposium on Cavitation (Osaka University, 159

S. R. Gonzalez-Avila, F. Denner, and C.-D. Ohl, “The acoustic pressure gener-
2003). ated by the cavitation bubble expansion and collapse near a rigid wall,” Phys.
156
K. V. Mani, A. Cervone, and J.-P. Hickey, “Turbulence modeling of cavitating Fluids 33, 032118 (2021).
flows in liquid rocket turbopumps,” J. Fluids Eng. 139, 011301 (2017). 160
T. Sun, Y. Ding, H. Huang, B. Xie, and G. Zhang, “Numerical study on the
157
Q. Liu, L. Gong, H. Gao, and R. K. Agarwal, “Numerical investigation of cavitation effects of modulated ventilation on unsteady cavity dynamics and noise pat-
characteristics of a liquid oxygen turbo pump,” AIAA Paper No. 2018-3222, 2018. terns,” Phys. Fluids 33, 123307 (2021).
158
Y. Fan, T. Chen, W. Liang, G. Wang, and B. Huang, “Numerical and theoreti- 161
A. Wei, S. Wang, X. Gao, L. Qiu, L. Yu, and X. Zhang, “Investigation of
cal investigations of the cavitation performance and instability for the cryo- unsteady cryogenic cavitating flow and induced noise around a three-
genic inducer,” Renewable Energy 184, 291 (2022). dimensional hydrofoil,” Phys. Fluids 34, 042120 (2022).
22 November 2024 04:35:04
Phys. Fluids 34, 101303 (2022); doi: 10.1063/5.0102876 34, 101303-40


Wei Review Cavitation PoF 2022

Uploaded by

Copyright:

Available Formats

Wei Review Cavitation PoF 2022

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

Wei Review Cavitation PoF 2022

Uploaded by

Copyright:

Available Formats

REVIEW ARTICLE | OCTOBER 21 2022

Cavitation in cryogenic fluids: A critical research review 

Physics of Fluids 34, 101303 (2022)

Articles You May Be Interested In

Cavitation flow characteristics on the surface of hydrofoil with microjet structure

Data-driven modal decomposition of transient cavitating flow

22 November 2024 04:35:04

Cavitation in cryogenic fluids: A critical research

Note: This paper is part of the special topic, Cavitation.

22 November 2024 04:35:04

NOMENCLATURE Pr Prandtl number

Phys. Fluids 34, 101303 (2022); doi: 10.1063/5.0102876 34, 101303-1

22 November 2024 04:35:04

FIG. 1. Typical engineering applications of cryogenic fluids.

Phys. Fluids 34, 101303 (2022); doi: 10.1063/5.0102876 34, 101303-2

22 November 2024 04:35:04

Phys. Fluids 34, 101303 (2022); doi: 10.1063/5.0102876 34, 101303-3

22 November 2024 04:35:04

Phys. Fluids 34, 101303 (2022); doi: 10.1063/5.0102876 34, 101303-4

22 November 2024 04:35:04

Phys. Fluids 34, 101303 (2022); doi: 10.1063/5.0102876 34, 101303-5

(Cf)r is expressed as follows at very high Reynolds numbers:

22 November 2024 04:35:04

Phys. Fluids 34, 101303 (2022); doi: 10.1063/5.0102876 34, 101303-6

Reference source of experimental data

Formulations of the B-factor Equation number Fluids Models Empirical parameters

GMR method Eq. (6) LH2 Venturi Velocity

22 November 2024 04:35:04

Phys. Fluids 34, 101303 (2022); doi: 10.1063/5.0102876 34, 101303-7

22 November 2024 04:35:04

FIG. 6. Schematic diagram of the experi-

Phys. Fluids 34, 101303 (2022); doi: 10.1063/5.0102876 34, 101303-8

22 November 2024 04:35:04

FIG. 7. Cavitation modes over inducer.

Phys. Fluids 34, 101303 (2022); doi: 10.1063/5.0102876 34, 101303-9

22 November 2024 04:35:04

2. Unsteady characteristics and flow instability B. Synchronous measurements

Phys. Fluids 34, 101303 (2022); doi: 10.1063/5.0102876 34, 101303-10

22 November 2024 04:35:04

Phys. Fluids 34, 101303 (2022); doi: 10.1063/5.0102876 34, 101303-11

TABLE III. Summary of experimental synchronous measurements of cryogenic cavitation.

Author Test sections Fluids Synchronous measurements

Ishimoto et al.50 Rectangular nozzle LN2 Visualization and pressure measurements

22 November 2024 04:35:04

Phys. Fluids 34, 101303 (2022); doi: 10.1063/5.0102876 34, 101303-12

TABLE IV. Common cavitation models.

Cavitation model _ þ (production term)

Merkle et al.71 Cprod Maxð0; p pv Þð1 al Þ Cdest ql Minð0; p pv Þal

22 November 2024 04:35:04

Phys. Fluids 34, 101303 (2022); doi: 10.1063/5.0102876 34, 101303-13

22 November 2024 04:35:04

Cavitation simulations Descriptions Comments

Phys. Fluids 34, 101303 (2022); doi: 10.1063/5.0102876 34, 101303-14

22 November 2024 04:35:04

Phys. Fluids 34, 101303 (2022); doi: 10.1063/5.0102876 34, 101303-15

22 November 2024 04:35:04

Phys. Fluids 34, 101303 (2022); doi: 10.1063/5.0102876 34, 101303-16

Cavitation p1 pmin p1 pmin

Kunz 71.08 1.22 40.01 0.86