Special Issue Article

Advances in Mechanical Engineering

2015, Vol. 7(11) 1–9
Ó The Author(s) 2015
Study of unsteady cavitation on DOI: 10.1177/1687814015618166
aime.sagepub.com
NACA66 hydrofoil using dynamic cubic
nonlinear subgrid-scale model

Xianbei Huang, Wei Yang, Zhuqing Liu and Yaojun Li

Abstract
In this article, we describe the use of a new dynamic cubic nonlinear model, a new nonlinear subgrid-scale model, for
simulating the cavitating flow around an NACA66 series hydrofoil. For comparison, the dynamic Smagorinsky model is
also used. It is found that the dynamic cubic nonlinear model can capture the turbulence spectrum, while the dynamic
Smagorinsky model fails. Both models reproduce the cavity growth/destabilization cycle, but the results of the dynamic
cubic nonlinear model are much smoother. The re-entrant jet is clearly captured by the models, and it is shown that the
re-entrant jet cuts the cavity into two parts. In general, the dynamic cubic nonlinear model provides improvement over
the dynamic Smagorinsky model for the calculation of cavitating flow.

Keywords
Large eddy simulation, nonlinear subgrid-scale model, hydrofoil, cavitation, re-entrant jet

Date received: 19 August 2015; accepted: 22 October 2015

Academic Editor: Moran Wang

Introduction
Cavitation is an unsteady flow phenomenon that usu-
ally has a negative or even destructive effect on hydrau-
lic machinery embodied in pressure pulsation, erosion,
and noise.1,2 Cavitation occurs when the liquid pressure
is lower than the saturated vapor pressure. In physical
experiments, cavitation is produced by decreasing the
pressure at the inlet or outlet of the flow field, leading
to different types of cavitation that can be observed
during this process. As the pressure decreases, an air
cavity is stably attached to the surface of blades, fol-
lowed by the development of cloud cavitation due to
the destabilization of the cavity, resulting in the shed-
ding of large vapor bubbles. Cloud cavitation is com-
monly associated with hydraulic machinery blades, and
it has a highly dynamic flow pattern.3 The intrinsic
mechanism is often attributed to the re-entrant jet gen-
erated at the end of the cavity.4–6
In order to simulate the unsteady characteristics of
cavitation through numerical calculation, it is usually
assumed that the vapor and liquid are mixed and that
they share the same velocity and pressure with a no-slip
velocity condition between the two phases, as proposed
by Kubota et al.7 Based on this assumption, Transport-
Based Equation Modeling (TEM) is widely used in
cavitating flows. With this method, an additional
advection equation is solved. Many cavitation models
have been used for this framework. For example,
Singhal et al.8 developed the full cavitation model;
Kunz et al.9 proposed a three-species formulation to
Beijing Engineering Research Center of Safety and Energy Saving
Technology for Water Supply Network System, College of Water
Resources & Civil Engineering, China Agricultural University, Beijing,
China
Corresponding author:
Wei Yang, Beijing Engineering Research Center of Safety and Energy
Saving Technology for Water Supply Network System, College of Water
Resources & Civil Engineering, China Agricultural University, 17 Qinghua
East Road, Haidian District, Beijing 100083, China.
Email: wyang@cau.edu.cn

Creative Commons CC-BY: This article is distributed under the terms of the Creative Commons Attribution 3.0 License
(http://www.creativecommons.org/licenses/by/3.0/) which permits any use, reproduction and distribution of the work without
further permission provided the original work is attributed as specified on the SAGE and Open Access pages (https://us.sagepub.com/en-us/nam/
open-access-at-sage).
Downloaded from ade.sagepub.com by guest on February 20, 2016
2 Advances in Mechanical Engineering

account for the liquid, vapor, and the non-condensable ∂r ∂ðrui Þ

+ =0 ð1Þ
gas. Apart from the advection equation, the other ∂t ∂xi
aspects related to TEM are the solution strategy for the  
∂rui ∂   ∂p ∂ ∂ui ∂t ij
advection equation and the turbulence model.10 + rui uj =  + m + ð2Þ
As a solution strategy, Volume of Fluid (VOF) ∂t ∂xj ∂xi ∂xj ∂xj ∂xj
method is a good technique with the advantage of the where the variables with a r bar over them represent the
ability to capture the cavity interface accurately. A filtered quantities, u is the velocity, p is the pressure, m
brief review of different VOF models can be found in is the laminar viscosity, and r is the mixture density
Roohi et al.10 The application of VOF in various cavi-
tating flows has been shown to be successful in tracking m = ð1  al Þmv + al ml ð3Þ
the cloud and super-cavitation cavities.10,11
For simulating turbulence, both Reynolds-averaged r = ð1  al Þrv + al rl ð4Þ
Navier–Stokes (RANS) and large eddy simulation where a is the volume fraction, and the subscripts v and
(LES) models are widely used, although it seems that l are the vapor and liquid, respectively. tij is the SGS
the results obtained with RANS modeling are not satis- stress, which is defined as
factory without modification of the original formula-
tions. For instance, Zhou and Wang12 studied a tij = ui uj  ui uj ð5Þ
two-dimensional (2D) hydrofoil and found that the
standard re-normalization group (RNG) k-e model
could not capture the cavity shedding unless the turbu-
DSM. DSM is a linear eddy viscosity model. The
lent viscosity was modified to account for cavitation.
Smagorinsky model can be expressed as
The theoretical basis of LES modeling is its ability
to capture large vortical structures including cloud cavi- 2 
tation. Liu et al.13 used the Smagorinsky–Lilly model tdij =  2Cs D S S ij ð6Þ
to study the cavitation on hydrofoil and successfully
where S ij is the rate-of-strain tensor, tij d is the deviato-
predicted the existence of the re-entrant jet; Ji et al.14
  of the SGS stress tensor, D is the filter width,
ric part
used the wall-adapting local eddy viscosity (WALE)
and S  = (2S ij S ij )1=2 is the characteristic filtered rate of
model and also reproduced the cavity shedding. Both
strain.
models needed no further modification, showing an
In order to dynamically calculate the model coeffi-
obvious advantage over the regular RANS models.
cient Cs , Germano18 proposed to apply a second filter,
However, the simulated pressure pulsation, which is an
called the test filter, to the SGS stress
important aspect of the cavitation process, tended to
fluctuate more than the experiment.14,15  
A review of the literature shows that the LES models Tij = e u j  ug
ui e i uj
ð7Þ
assume a linear relationship between the subgrid-scale Lij = Tij  e
t ij
(SGS) stress and the rate-of-strain, which may be the rea-
son for the disagreement with experiment, as it has been where ‘‘;’’ is the test-filtering operation and Tij is the
figured out that the eigenvectors of the strain rate tensor e = 2D). Similar to equation
stress at a test filter scale ( D
and SGS stress tensor are not aligned but with a certain (6)
degree between each other in this linear relation.16,17 A
more reasonable relation to use would be the nonlinear e 2 f
Tijd =  2Cs D

S e
S ij ð8Þ
tensorial polynomial constitutive form, which combines
SGS stress with strain rate and vorticity. where Tij d is the deviatoric part of Tij . Combining equa-
In order to examine the performance of nonlinear tions (6) and (8)
models in simulating cavitating flow, a new nonlinear
2  e 2 f
SGS model is proposed to simulate the cavitation around t dij = 2Cs D S S ij  2Cs D
Tijd  e S e
S ij = Cs Mij ð9Þ
a hydrofoil. The results are compared with the experimen-
tal data obtained by Leroux et al.6 and with the numerical Using equation (9) to provide an approximation to Ldij
results of the dynamic Smagorinsky model (DSM).
Ldij = Cs Mij ð10Þ
Governing equations
The use of least squares minimization of the error is
LES modeling suggested by Lilly19 for the calculation of the value of Cs
In the LES model, a low-pass filter (in frequency) is
applied to the Navier–Stokes equations to get the fol- Mij Lij
Cs = ð11Þ
lowing formulations (in Cartesian coordinates) Mkl Mkl

Downloaded from ade.sagepub.com by guest on February 20, 2016

Huang et al. 3

   

It is well known that DSM suffers from numerical e

2 e O e e e e 1e e
Nij = 4D S ik O kj S
ik kj  S S
ik kj  S S d
mn nm ij
instability due to the large oscillation of the model coef- 3
ficient. In this article, the eddy viscosity is bounded as    

1 g
Sg g g
2
D O
ik kj  O S
ik kj  S S
ik kj  S S d
mn nm ij
3
nr = maxðnr ,  nÞ ð12Þ
ð18Þ
which means that the backscatter cannot be larger than 2
the viscous dissipation (n is the kinetic viscosity). Pij = 4D
2 3
 
1
4  O e e 2 e e
2 e e e 1 e e e 1 e e 5
2
ik S kj  S ik O kj + O ik S km O mj  O kl S lm O mk dij  S kk S ij
f
S 3 2
Dynamic cubic nonlinear model. Kosovic20 proposed "  #
applying the concepts of nonlinear models of different 2 1 2 2 1 1 2
 D   (Oik S kj  S ik Okj + Oik S km Omj  Okl S lm Omk dij  S kk S ij
levels of complexity for Reynolds stress closures to S  3 2
SGS model as they have been extensively studied by ð19Þ
researchers and are much more developed. Therefore,
Shih et al.’s method21 is applied for the construction of We named the dynamic cubic nonlinear model
the new model. The general form of the new model can (DCNM) as the new model contains a cubic term. In
be written as follows order to guarantee the numerical stability and save
computation resources, the coefficients in equation (16)
2  2  2
tdij = Cs D S S ij + C1 D S ik Okj  Oik S kj + C2 D are expressed explicitly with the volume average applied
" #
1 2 2 1 to both the numerator and denominator of each coeffi-
  (Oik S kj  S ik Okj + Oik S km Omj  Okl S lm Omk dij + Ps S ij cient’s expression.
S  3

ð13Þ
Two-phase modeling
where Ps is the second principal invariant of Sij
As discussed above, TEM is used and the mixture
1 2
 model is considered in this article. In the mixture
Ps = Skk Smm  Skk ð14Þ
2 model, the fluids are treated as a combination of vapor
and liquid, and an additional transport equation is
Adapting the above model to the simple shear flow,
written as
it can be shown that equation (13) does not stand.
Comparing with the Kosovic model,20 the term ∂al ∂ðal ui Þ
(S ik S kj  (1=3)S mn S nm dij ) is introduced into the new + = m_ ð20Þ
∂t ∂xi
model and the final form appears as
where al is the volume fraction of liquid and m_ is the
2  2
tdij = Cs D S S ij + C1 D mass transfer rate between the two phases. To express
 
the mass transfer rate, the Kunz et al. model9 is
  1 2
S ik Okj  Oik S kj  S ik S kj  S mn S nm dij + C2 D adopted
3
"  #
1 2 2 1 Cdest rv min½0, p  pv al
  Oik S kj  S ik Okj + Oik S km Omj  Okl S lm Omk dij + Ps S ij
S  3 m_  =   ð21Þ
rl 0:5rl U‘2 t‘
ð15Þ
Cprod ð1  al Þa2l
In order to dynamically calculate the three model m_ + = ð22Þ
r l t‘
coefficients (Cs, C1, and C2), the method proposed by
Wang and Bergstrom22 is applied. This is an extension where m_  and m_ + represent condensation and evapora-
of Germano’s method. The least square method is used tion, respectively; Cdest and Cprod are empirical con-
stants; U‘ is the free-stream velocity; t‘ is the flow
2 32 3 2 3
Mij Mij Mij Nij Mij Pij Cs Mij Ldij characteristic time defined as t‘ = L=U‘ , where L is the
4 Nij Mij 6 7
Nij Nij Nij Pij 54 C1 5 = 4 Nij Ldij 5 ð16Þ body diameter; and pv is the saturation pressure.
Pij Mij Pij Nij Pij Pij C2 d
Pij Lij In OpenFOAM,23 the VOF method is an improved
version of CICSAM (Compressive Interface Capturing
where Scheme for Arbitrary Meshes) as follows

2 e 2  ∂al ∂ðal ui Þ ∂½Ui al ð1  al Þ

Mij = 4D f
S  S ij  D Sg
S ij ð17Þ + + =0 ð23Þ
∂t ∂xi ∂xi

Downloaded from ade.sagepub.com by guest on February 20, 2016

4 Advances in Mechanical Engineering

Figure 1. Computational domain and boundary conditions.

The compression velocity is calculated as algorithm is used for the pressure–velocity coupling,
and the time-step size is set to 1 3 105 s to ensure that
ral the maximum Courant number Co\1.
U = min½Cl jUj, maxðjujÞ ð24Þ
jral j The pressure coefficient is defined as
where Cl is a constant which controls the compression. pm  pout
The surface tension is evaluated by Cp = ð28Þ
0:5rU‘2
fs = skral ð25Þ where pm is the time-averaged (8 3 104 time-steps are
used for averaging) pressure at the monitor points,
where s is the surface tension coefficient, and the free
which are equally spaced on the suction side of foil
surface curvature k is defined as
within 0:1c  0:9c from the leading edge; pout is the
  pressure at outlet; and U‘ = 5:33 m=s is the free-stream
ral
k=  r  ð26Þ velocity, which is set to be the inlet velocity. Both the
jral j
fine-grid convergence index and the extrapolated value
The surface tension is added to the momentum equa- for Cp are shown in Figure 2.
tion, and thus equation (2) is rewritten as From the error bar shown in Figure 2(a), it can be
  seen that the fine-grid convergence index of the grid is
∂rui ∂   ∂p ∂ ∂ui ∂t ij small with an acceptable maximum error below 0.07.
+ rui uj =  + m + + fs
∂t ∂xj ∂xi ∂xj ∂xj ∂xj Figure 2(b) shows that the extrapolated value is close to
that of the fine grid (with the number of grid elements
ð27Þ
equal to 5:1 3 105 ). Hence, the fine grid is adopted for
the modeling.
Numerical setup
The hydrofoil investigated in this study is an NACA66 Results and discussions
series foil fixed in a square channel, with the chord
length c = 0:15 m and span width 0:19 m. The specific According to the experimental results reported by
parameters of the foil can be found in Leroux et al.6 In Leroux et al.,6 when the cavitation number defined as
order to reduce the calculation time, the span width is s = (pout  psat )=(0:5rU‘2 ) (psat is the saturation pres-
set to 0:3c, with the calculation domain as illustrated in sure) reaches 1.25, the unsteady cavitating phenomenon
Figure 1. will occur with the cavity periodically growing, shed-
The procedure proposed by Celik et al.24 is used to ding, and disappearing. This unsteady behavior will
examine whether the grid is in the asymptotic range. claim more requirements on the accuracy of the turbu-
To resolve the near-wall characteristics, y+ is set to be lence model. In the following calculations, DSM and
located in the viscous sub-layer at walls. Three sets of DCNM are used under the cavitating condition of
grids with 8:8 3 104 , 2:2 3 105 , and 5:1 3 105 cells, s = 1:25.
respectively, are generated. DSM and the Kunz et al.
model9 are used to simulate the hydrofoil for the non-
cavitating condition. The open-source code
Pressure coefficient and cavity shedding frequency
OpenFOAM23 is used for all the calculations. As stated Figure 3 shows a plot of the time-averaged value of Cp
above, a mixture method is employed for the two-phase calculated along the suction side of the hydrofoil with
modeling, and VOF is adopted to solve the additional the DSM and DCNM methods and the referenced
transfer equation in the cavitation model. The PISO experimental data. It can be seen that the pressure

Downloaded from ade.sagepub.com by guest on February 20, 2016

Huang et al. 5

Figure 2. (a) Fine-grid convergence index and (b) extrapolated value comparing to the result of fine grid.

clearly explained by the instantaneous pressure. When

Figure 3. Comparison of the pressure coefficients along the
suction side.
coefficients are overestimated at x=c\0:5 and underes-
timated at x=c.0:5. This tendency is similar to the
results obtained by Zhou and Wang,12 which can be
x=c\0:5, the predicted pressure fluctuation shows a
shorter duration of low pressure (as shown in Figure 6)
which leads to the overestimation; when x=c.0:5, the
duration of high pressure is shorter (not shown), lead-
ing to the underestimation of the pressure coefficients.
Despite the discrepancy with the experimental data,
DCNM shows better agreement.
Due to the intrinsic relationship between cavitation
and pressure, it is convenient to determine the cavity
shedding frequency from the pressure signal. As was
done in Ji et al.,14 the power spectral density (PSD) is
used in this article. Figure 4 shows the PSD profiles of
two models, obtained in both cases by taking 7:3 3 104
samples at x=c = 0:6 on the foil suction side. For both
models, the cavity shedding frequency is 4.1 Hz, which
is higher than the experimental value 3.625 Hz. In order
to examine the models’ ability to capture the turbulence
spectrum, the 25/3 slope is shown as an indication of
the model accuracy.14 It is clear that DCNM captures

Figure 4. PSD of DSM and DCNM at x=c = 0:6 on the suction side: (a) DSM and (b) DCNM.

Downloaded from ade.sagepub.com by guest on February 20, 2016

6 Advances in Mechanical Engineering

Figure 5. Cavity evolution in one cycle (time interval between two images: numerical data, 0.05 s; experimental data, 0.08 s): (a)
DSM, (b) DCNM and (c) experiment.

Figure 6. Comparison of instantaneous pressure at x=c = 0:4 on the suction side: (a) DSM and (b) DCNM.

this slope accurately, while DSM fails, showing a much
more flat tendency.
Cavity growth/destabilization cycle
The cavity growth/destabilization cycle can be clearly
illustrated by the numerical results, as shown in
Figure 5. The iso-surface of the vapor volume fraction
of 0.1 is used to show the cavity shape. It is observed
that the cavity basically experiences three stages: (1) the
cavity occurs at the leading edge and develops down-
stream; (2) when the cavity length is maximum, the cav-
ity is cut into a partial cavity and a cloud cavitation by
a pressure perturbation (shown in Figure 6); and (3)

Downloaded from ade.sagepub.com by guest on February 20, 2016

Huang et al.
Figure 7. Predicted re-entrant flow on the suction side: (a)
DSM and (b) DCNM.
when the cloud cavitation reaches the trailing edge, the
cavitation suddenly disappears (not shown). Both the
models correctly reproduce the whole cycle compared
with the experimental results. It should be noted that
three-dimensional characteristics are shown in the
spanwise direction, which would probably lead to the
overestimation of the shedding frequency due to the
reduced span width used in the calculations.
Three stages can be reflected in the temporal pres-
sure signal on the suction side of the hydrofoil. Figure 6
shows a comparison of instantaneous pressure signals
at x=c = 0:4 on the suction side. In order to compare
these values with the experimental data, it is necessary
to choose a base point. Leroux et al.6 found the major
pressure perturbation occurs at approximately
t = 0:411 s and cuts the cavity into two parts (stage 2 in
Figure 5). Therefore, the points when the cavity is cut
into two parts in DSM and DCNM are found accord-
ingly and matched to t = 0:411 s. On the whole, the two
models perform similarly: at t = 0:32  0:64 s, the
results agree well with the experimental data and the
major pressure perturbation is reproduced around
t = 0:411 s, while at t = 0:2  0:32 s, the peak value of
the pressure moves approximately 0.0645 s (for numeri-
cal results, the peak value is at approximately
t = 0:31 s; for experimental data, it is around
t = 0:245 s), which is equal to twice the phase difference
between the numerical results and experimental data.
This results from the step matching one cycle
(t = 0:32  0:64 s) to the experimental data, and then
the phase difference of two cycles is imposed on the pre-
vious cycle (t = 0  0:32 s). The major difference
between the two results is that the data of DCNM tend
to be much smoother than those of DSM. This may be
the reason why DSM cannot reflect the right turbulence
spectrum in Figure 4.
At stage 2, the cavity shedding is attributed to
the re-entrant flow generated at the end of the
Downloaded from ade.sagepub.com by guest on February 20, 2016
7
Figure 8. Comparison of the instantaneous non-dimensional
model coefficient at x=c = 0:4.
cavity.4–6 Figure 7 shows the velocity vector field
around the foil at t = 0:411 s when the cavity is cut into
two parts. In both images, the re-entrant flow is pre-
dicted at approximately x=c = 0:6, where the flow is
cut by it. It should be noted that the re-entrant flow
predicted by DSM is some distance away from the foil
surface. Due to the pressure gradient, part of the main
flow reverses direction and cuts the cavity from the bot-
tom.5 Therefore, the re-entrant flow should be exactly
on the suction side as shown in Figure 7(b). This
should be attributed to the insufficiency of DSM in the
near-wall behavior, which is also documented in Yang
et al.25
Model coefficients and SGS dissipation
From the above analysis, it is found that the instanta-
neous pressure of DCNM is smoother than the DSM’s.
The underlying reason can be attributed to the model
behavior during the calculation. The model coefficient
is a representative quantity which will vary during the
calculation. Figure 8 shows the instantaneous non-
dimensional model coefficient, which is defined as
DC = (C  mean(C))=(mean(C)) at x=c = 0:4, where C
represents the model coefficients of the models—Cs in
DSM and Cs , C1 , C2 in DCNM.
As can be seen, the fluctuation of the model coeffi-
cient in DSM is much larger than in DCNM. This is a
good indication that the pressure field calculated by
DSM experiences more fluctuation.
Another index quantity is the non-dimensional SGS
turbulent kinetic energy (TKE) production, which is
defined as Pr =  tdij S ij =(U‘4 =n), representing the TKE
transferred to the residual motions. Figure 9 shows a
comparison of the instantaneous SGS TKE production
for the two models at x=c = 0:4. Most of the time, the
SGS TKE production of DSM is negative, which means
that the energy is transferred from the residual motions
to the large-scale motions, namely, the backscatter.
8 Advances in Mechanical Engineering
Figure 9. Comparison of the instantaneous non-dimensional
SGS TKE production at x=c = 0:4.
Therefore, the large-scale motions, including the pres-
sure field fluctuation, tend to be influenced by the sub-
grid scale motions. Additionally, the fluctuation of Pr
in DSM is also much larger than in DCNM, having an
even larger effect on the large-scale motions.
Conclusion
In this article, we present the use of a linear model
DSM and the new nonlinear model DCNM for the cal-
culation of cavitating flow around an NACA66 series
hydrofoil and the simulation results are compared with
the experimental data. For both models, the cavity
shedding frequency is overestimated, but the reason is
still not clear. It is suspected that the reduced span
width in the simulation may affect the shedding fre-
quency as the cavitating flow shows three-dimensional
characteristics. From the PSD analysis, it is found that
DCNM can capture the turbulence spectrum while
DSM fails to do so. The cavity growth/destabilization
cycle, simulated by both models, duplicates the experi-
mental data with little discrepancy.
Further investigation is done to explore the instanta-
neous pressure on the suction side. The results of
DCNM are much smoother than in DSM, and this
phenomenon may explain why DSM fails to capture
the turbulence spectrum. Additionally, the re-entrant
jet which cut the cavity into two parts is captured by
the models. The instantaneous model coefficients and
SGS TKE production are studied to reveal the underly-
ing reason why DCNM predicts smoother pressure
profile. It is found that both the model coefficient fluc-
tuation and SGS TKE production are much larger for
DSM. Additionally, significant backscatter of DSM is
observed, which will affect the large-scale motions.
Under these conditions, the pressure field is influenced
in DSM, resulting in much fluctuation of the instanta-
neous pressure. In general, the nonlinear model does
improve the ability to capture the turbulence, while it
Downloaded from ade.sagepub.com by guest on February 20, 2016
should be noted that the cavity shedding frequency is
turbulence model independent. Further improvements
of the calculation may focus on the cavitation model.
Declaration of conflicting interests
The author(s) declared no potential conflicts of interest with
respect to the research, authorship, and/or publication of this
article.
Funding
The author(s) disclosed receipt of the following financial sup-
port for the research, authorship, and/or publication of this
article: The authors would like to acknowledge the financial
support given by the Ministry of Education, Science and
Technology Project (Grant No. 113010A), the Research Fund
for the Doctoral Program of Higher Education of China
(Grant No. 20130008110047), and the National Natural
Science Foundation of China (Grant No. 51209206).
References
1. Zhang YN and Li SC. A general approach for rectified
mass diffusion of gas bubbles in liquids under acoustic
excitation. J Heat Trans: T ASME 2014; 136: 042001.
2. Roohi E, Pendar MR and Rahimi A. Simulation of
three-dimensional cavitation behind a disk using various
turbulence and mass transfer models. Appl Math Model.
Epub ahead of print 2 July 2015. DOI: 10.1016/
j.ijheatmasstransfer.2013.10.019.
3. Arndt REA. Some remarks on hydrofoil cavitation. J
Hydrodyn 2012; 24: 305–314.
4. Pham TM, Larrarte F and Fruman DH. Investigation of
unstable sheet cavitation and cloud cavitation mechan-
isms. J Fluid Eng: T ASME 1999; 121: 289–296.
5. Callenaere M, Franc JP, Michel JM, et al. The cavitation
instability induced by the development of a re-entrant
jet. J Fluid Mech 2001; 444: 223–256.
6. Leroux JB, Astolfi JA and Billard JY. An experimental
study of unsteady partial cavitation. J Fluid Eng: T
ASME 2004; 126: 94–101.
7. Kubota A, Kato H and Yamaguti H. A new modeling of
cavitating flows: a numerical study of unsteady cavita-
tion on a hydrofoil section. J Fluid Mech 1992; 240:
59–96.
8. Singhal NH, Athavale AK, Li M, et al. Mathematical
basis and validation of the full cavitation model. J Fluid
Eng: T SAME 2002; 124: 1–8.
9. Kunz RF, Boger DA, Stinebring DR, et al. A precondi-
tioned Navier–Stokes method for two-phase flows with
application to cavitation. Comput Fluids 2000; 29:
849–875.
10. Roohi E, Zahiri AP and Passandideh-Fard M. Numeri-
cal simulation of cavitation around a two-dimensional
hydrofoil using VOF method and LES turbulence model.
Appl Math Model 2013; 37: 6469–6488.
11. Passandideh-Fard M and Roohi E. Transient simulations
of cavitating flows using a modified volume-of-fluid
(VOF) technique. Int J Comput Fluid D 2008; 22: 97–114.
Huang et al.
12. Zhou LJ and Wang ZW. Numerical simulation of cavita-
tion around a hydrofoil and evaluation of a RNG k-e
model. J Fluid Eng: T ASME 2008; 130: 011302.
13. Liu DM, Liu SH, Wu YL, et al. LES numerical simula-
tion of cavitation bubble shedding on ALE 25 and ALE
15 hydrofoils. J Hydrodyn 2009; 21: 807–813.
14. Ji B, Luo XW, Arndt REA, et al. Large Eddy Simulation
and theoretical investigations of the transient cavitating
vortical flow structure around a NACA66 hydrofoil. Int
J Multiphas Flow 2015; 68: 121–134.
15. Hidalgo VH, Luo XW, Escaler X, et al. Numerical inves-
tigation of unsteady cavitation around a NACA 66
hydrofoil using OpenFOAM. In: 27th IAHR symposium
on hydraulic machinery and systems, Montreal, QC,
Canada, 22–26 September 2014. IOP Publishing.
16. Tao B, Katz J and Meneveau C. Statistical geometry of
subgrid-scale stresses determined from holographic parti-
cle image velocimetry measurements. J Fluid Mech 2002;
457: 35–78.
17. Horiuti K. Roles of non-aligned eigenvectors of strain-
rate and subgrid-scale stress tensors in turbulence genera-
tion. J Fluid Mech 2003; 491: 65–100.
18. Germano M. Turbulence: the filtering approach. J Fluid
Mech 1992; 238: 325–336.
Downloaded from ade.sagepub.com by guest on February 20, 2016
9
19. Lilly DK. A proposed modification of the Germano
subgrid-scale closure method. Phys Fluids A: Fluid 1992;
4: 633–635.
20. Kosovic B. Subgrid-scale modelling for the large-eddy
simulation of high-Reynolds-number boundary layers. J
Fluid Mech 1997; 336: 151–182.
21. Shih TH, Zhu J and Liou W. Modeling of turbulent swir-
ling flows. NASA technical memorandum 113112, 1997,
http://ntrs.nasa.gov/archive/nasa/casi.ntrs.nasa.gov/
19970034949.pdf
22. Wang BC and Bergstrom DJ. A dynamic nonlinear
subgrid-scale stress model. Phys Fluids A: Fluid 2005; 17:
035109-1–035109-15.
23. OpenCFD. OpenFOAMÒ, The opens source CFD toolbox,
user guide. Berkshire: OpenCFD Ltd, 2012.
24. Celik IS, Ghia U, Roache PJ, et al. Procedure for estima-
tion and reporting of uncertainty due to discretization in
CFD applications. J Fluid Eng: T ASME 2008; 130:
078001.
25. Yang ZX, Cui GX, Xu CX, et al. Large eddy simulation
of rotating turbulent channel flow with a new dynamic
global-coefficient nonlinear subgrid stress model. J
Turbul 2012; 13: 1–20.