You are on page 1of 19

Nuclear Engineering and Design 230 (2004) 233251

A review of heat exchanger tube bundle vibrations


in two-phase cross-ow
Shahab Khushnood

, Zaffar M. Khan, M. Afzaal Malik,


Zafar Ullah Koreshi, Mahmood Anwar Khan
College of Electrical & Mechanical Engineering, National University of Sciences and Technology, Rawalpindi, Pakistan
Received 8 May 2003; received in revised form 8 October 2003; accepted 21 November 2003
Abstract
Flow-induced vibration is an important concern to the designers of heat exchangers subjected to high ows of gases or liquids.
Two-phase cross-ow occurs in industrial heat exchangers, such as nuclear steam generators, condensers, and boilers, etc. Under
certain owregimes and uid velocities, the uid forces result in tube vibration and damage due to fretting and fatigue. Prediction
of these forces requires an understanding of the ow regimes found in heat exchanger tube bundles. Excessive vibrations under
normal operating conditions can lead to tube failure.
Relatively little information exists on two-phase vibration. This is not surprising as single-phase ow induced vibration; a
simpler topic is not yet fully understood. Vibration in two-phase is much more complex because it depends upon two-phase ow
regime, i.e. characteristics of two-phase mixture and involves an important consideration, which is the void fraction. The effect
of characteristics of two-phase mixture on ow-induced vibration is still largely unknown. Two-phase ow experiments are
much more expensive and difcult to carry out as they usually require pressurized loops with the ability to produce two-phase
mixtures. Although convenient from an experimental point of view, airwater mixture if used as a simulation uid, is quite
different from high-pressure steamwater. A reasonable compromise between experimental convenience and simulation of
steamwater two-phase ow is desired.
This paper reviews known models and experimental research on two-phase cross-ow induced vibration in tube bundles.
Despite the considerable differences in the models, there is some agreement in the general conclusions. The effect of tube bundle
geometry, random turbulence excitations, hydrodynamic mass and damping ratio on tube response has also been reviewed.
Fluidstructure interaction, void fraction modeling/measurements and nally Tubular Exchanger Manufacturers Association
(TEMA) considerations have also been highlighted.
2004 Elsevier B.V. All rights reserved.
1. Introduction
Two-phase cross-ow induced vibration in tube
bundles of process heat exchangers and U-bend region
of nuclear steam generators can cause serious tube

Corresponding author. Tel.: +92-51-2873423;


fax: +92-51-2824132.
E-mail address: seeshahab@yahoo.com (S. Khushnood).
failures by fatigue and fretting wear. Tube failures
could force entire plant to shutdown for costly repairs
and suffer loss of production. Such vibration prob-
lems may be avoided by thorough vibration analysis.
However, these require an understanding of vibra-
tion excitation and damping mechanism in two-phase
ow. An important parameter that characterizes the
two-phase ow is void fraction, which is the ratio of
the volume of gas to the volume of the liquidgas
0029-5493/$ see front matter 2004 Elsevier B.V. All rights reserved.
doi:10.1016/j.nucengdes.2003.11.024
234 S. Khushnood et al. / Nuclear Engineering and Design 230 (2004) 233251
Nomenclature
a tube gap
A
v
upstream ow area
b exponent
B Chisholm correlation parameter
C connement factor
C
1
, C
0
drift ux model parameters
C
I
coefcient of interaction
C

1
non-dimensional liquid force
coefcient
C

g
non-dimensional gas force
coefcient
Cap capillary number
d, D tube diameter
D
e
equivalent diameter
Eu Euler number
Eu
G0
Euler number for gas
Eu
L0
Euler number for liquid
f tube natural frequency in two-phase
mixture
f
g
tube natural frequency in air
F instantaneous normal force
Fr Schrage correlation parameter
FIV ow induced vibration
g acceleration due to gravity
G
P
pitch mass ux, mass ow
HEM Homogeneous Equilibrium Model
i mode number
j
g
supercial gas velocity
j
l
supercial liquid velocity
K constant fraction in Smith correlation,
Connors critical factor
K
0
, K
1
drift ux model parameters
L, l tube length, limit span subjected to
cross-ow
L drift ux model parameter
m mass, mass per unit length
m
0
mass
m
geq
gas phase effective mass
m
h
hydrodynamic mass/added mass
m
leq
liquid phase effective mass
m
t
mass of tube alone
m mass ow rate
m
p
mass ux of mixture
n number of points discretized/index in
Chisholm correlation
N gamma count of experimental (trial)
N
G
gamma count for 100% gas
N
L
gamma count for 100% liquid
NPSD Normalized Power Spectral Density
p
D
dynamic pressure
p
S
static pressure
P pressure, tube pitch
P
cr
two-phase ow critical pressure
p pressure drop across tube row
p
G0
Pressure drop for gas
p
L0
Pressure drop for liquid
p
TPF
two-phase ow pressure drop
PDF probability density function
Q mass ux
r drift ux model parameter
Re Reynolds number
Ri Richardson number
RAD Radiation Attenuation Method
R-11, R-12 refrigerant types
RMS root mean square
S velocity ratio/slip
S
F
(f) power spectral density of random
excitation force
S sliding distance
T time period
T
s
sample duration
TEMA Tubular Exchanger Manufacturers
Association
U
g
, U
G
gas velocity
U
l
, U
L
liquid velocity
U
h
homogeneous velocity
U
p
, V
p
pitch velocity
U
c
critical velocity
U
cl
critical ow velocity of liquid phase
U
cg
critical ow velocity of gas phase
V
eq
equivalent two-phase velocity
U
gj
averaged gas phase drift velocity
W work rate
x location
X quality of ow/mass quality
Y Chisholm parameter
Y(1),
Y(x), Y tube response
(Y
2
) mean square tube response
(Y
2
)
0.5
root mean square tube response
S. Khushnood et al. / Nuclear Engineering and Design 230 (2004) 233251 235
Greek symbols
void fraction
average value of (s)T
(time length of gas phase)

H
homogeneous void fraction
non-dimensional parameter
non-dimensional parameter
average damping in two-phase

g
homogeneous void fraction
total damping ratio, modal damping
ratio

eq
non-dimensional effective ow
velocity of gas phase
ratio of non-dimensional effective
masses per unit length of tube in
liquid and gas phases

L
liquid phase absolute viscosity
viscosity of mixture in McAdams
equation

l
viscosity of liquid phase

g
viscosity of gas phase

eq
non-dimensional effective ow
velocity of liquid phase
two-phase mixture density

g
,
G
gas mass density

l
,
L
liquid phase density

H
HEM uid density
density difference between phases
liquid surface tension

2
L0
two-phase friction multiplier

L0
function of two-phase ow
parameters

i
(S) shape function
(S) ow distribution
Table 1
Types of ow in two-phases
Flow type Average void fraction Specication
Bubble 0.3 Some bubbles are present in liquid ow and move with the same velocity
Slug 0.30.5 Liquid slugs ow intermittently
Froth 0.50.8 More violent intermittent ow
Annular 0.80.9 Mainly gas ow, liquid adheres to the tube surface
Mist 0.9 Almost gas ow. Mist sometimes causes energy dissipation
mixture. A number of ow regimes (Table 1) can
occur for a given boundary conguration, depending
upon the concentration and size of the gas bubbles and
on the mass ow rates of the two-phases. Two-phase
ow characteristics greatly depend upon the type of
ow occurring.
Tube vibration in two-phase ow displays different
ow regimes, i.e. gas and liquid phase distributions,
depending upon the void fraction and mass ux. It is
known that four mechanisms are responsible for the
excitation of tube arrays in cross-ow (Pettigrew et al.,
1991). These mechanisms are turbulence buffeting,
vortex shedding or Strouhal periodicity, uid-elastic
instability and acoustic resonance. Table 2 presents
a summary of these vibration mechanisms for sin-
gle cylinder and tube bundles for liquid, gas and
liquidgas two-phase ow, respectively. Of these four
mechanism, uid-elastic instability is the most dam-
aging in the short term because it causes the tubes to
vibrate excessively, leading to rapid wear at the tube
supports. This mechanism occurs once the ow rate
exceeds a threshold velocity at which tubes become
self excited and the vibration amplitude rise rapidly
with an increase in ow velocity.
Typically, researchers have relied on the Homo-
geneous Equilibrium Model (HEM) to dene impor-
tant uid parameters in two-phase ow, such as den-
sity, void fraction and velocity. This model treats the
two-phase ow as a nely mixed and homogeneous
in density and temperature, with no difference in ve-
locity between the gas and liquid phases. This model
has been used a great deal because it is easy to im-
plement and is widely recognized which facilitated
earlier data comparison. Other models include Smith
Correlation (Smith, 1968), drift-ux model developed
by Zuber and Findlay (1965), Schrage Correlation
(Schrage, 1988) which is based on empirical data, and
Feenstra model (Feenstra et al., 2000) which is given
in terms of dimensionless numbers.
236 S. Khushnood et al. / Nuclear Engineering and Design 230 (2004) 233251
Table 2
Vibration excitation mechanisms (Pettigrew et al., 1991)
Flow situation Fluid-elastic instability Periodic shedding Turbulence excitation Acoustic resonance
Cross ow
Single cylinder
Liquid
a

a
Gas
a

a
Two phase
a

a
Tube bundle
Liquid
b

a
Gas
b

b
Two phase
b

a
a
Unlikely.
b
Possible.
c
Most important.
Dynamic parameters such as added mass and damp-
ing are very important considerations in two-phase
cross-ow induced vibrations. Hydrodynamic mass
depends upon pitch-to-diameter ratio and decrease
with increase in void fraction. Damping is very com-
plicated in two-phase ow and is highly void fraction
dependent.
Tube-to-restraint interaction at the bafes (loose
supports) can lead to fretting wear because of out of
plane impact force and in-plane rubbing force. Frick
et al. (1984) has given an overview of the development
of relationship between work-rate and wear-rate.
Another important consideration in two-phase ow
is the random turbulence excitation. Vibration re-
sponse below uid-elastic instability is attributed to
random turbulence excitation (Pettigrew et al., 2000;
Mirza and Gorman, 1973; Taylor et al., 1989; Papp,
1988; Wambsganss et al., 1992) to name some who
have carried potential research for RMS vibration re-
sponse encompassing spatially correlated forces, Nor-
malized Power Spectral Density (NPSD), two-phase
ow pressure drop, two-phase friction multiplier,
mass ux, and coefcient of interaction between uid
mixture and tubes.
Earlier reviews on the topic are provided by
Paidoussis (1982), Weaver and FitzPatrick (1988),
and Price (1995). More recently researchers have
expanded the study to two-phase ow which occur
in nuclear steam generators and many other tubular
heat exchangers, a review of which was last given by
Pettigrew and Taylor (1994). The aim of present at-
tempt is to review all known models and experimental
research on two-phase cross-ow induced vibrations
of tube bundles. It is intended to provide design
guidelines to the heat exchanger designers, to give an
insight to the process designers and the maintenance
personnel.
2. Modeling two-phase ow
Most of the early experimental research in this eld
relied on sectional models of tube arrays subjected to
single-phase uids such as air or water, using rela-
tively inexpensive ow loops and wind tunnels. The
cheapest and simplest approach to model two-phase
ow is by mixing air and water at atmospheric pres-
sure. However, airwater ows have a different den-
sity ratio between phases than steamwater ow and
this will affect the difference in the ow velocity be-
tween the phases. The liquid surface tension, which
controls the bubble size, is also not accurately modeled
in airwater mixtures. Table 3 gives the comparison
of liquid and gas phase of refrigerants R-11, R-22 and
airwater mixtures at representative laboratory con-
ditions with actual steamwater mixture properties at
typical power plant conditions (Feenstra et al., 2000).
This comparison reveals that the refrigerants approxi-
mate the liquid surface tension and liquid dynamic vis-
cosity of steamwater mixtures more accurately than
airwater mixtures.
In the outer U-bend region, of typical nuclear steam
generators such as those used in the CANDU de-
sign, the tubes are subjected to two-phase cross-ow
S. Khushnood et al. / Nuclear Engineering and Design 230 (2004) 233251 237
Table 3
Comparison of properties of airwater, R-22, and R-11 with steamwater at plant conditions (Feenstra et al., 2000)
Property R-11 Airwater R-22 Steamwater
Temperature (

C) 40 22 23.3 260
Pressure (kPa) 175 101 1000 4690
Liquid density (kg/m
3
) 1440 998 1197 784
Gas density (kg/m
3
) 9.7 1.18 42.3 23.7
Liquid kinematic viscosity (m
2
/s) 0.25 1.0 0.14 0.13
Gas kinematic viscosity (m
2
/s) 1.2 1.47 0.30 0.75
Liquid surface tension (N/m) 0.016 0.073 0.0074 0.0238
Density ratio 148 845 28.3 33
Viscosity ratio 0.20 0.70 0.47 0.17
of steamwater. It is highly impractical and costly
to perform ow induced vibration experiments on
a full-scale prototype of such a device so that
small-scale sectional modeling is most often adopted.
R-11 simulates the density ratio, viscosity ratio and
surface tension of actual steamwater mixtures better
than airwater mixtures and it also allows for local-
ized phase change which airwater mixture does not
permit. While more costly and difcult to use than
airwater mixture, R-11 is a much cheaper modeling
uid than steamwater because it requires 8% of the
energy to evaporate the liquid and operating pressure
is much lower, thereby reducing the size and cost of
the ow loop (Feenstra et al., 2000).
3. Representative published tests on two-phase
ow across tube arrays
Table 4, an extension of period beyond 1993
(Nakamura et al., 1993) presents summary of salient
features of the experimental tests performed on the
three possible geometric tube arrangements.
4. Thermal hydraulic models
Considering two-phase ow, homogeneous ow as-
sume that the gas and liquid phases are owing at the
same velocity, while other models for two-phases ow,
such as drift-ux assume a separated ow model with
the phases allowed to ow at different velocities. Gen-
erally the vapor ow faster in upward ow because of
the density difference.
4.1. The homogeneous equilibrium model (HEM)
A general expression for void fraction , is given
by (Feenstra et al., 2000):
=
_
1 +S

L
_
1
X
1
__
1
(1)
where
G
and
L
are the gas and liquid densities, re-
spectively and S is the velocity ratio of the gas and liq-
uid phase (i.e. S = U
G
/U
L
). The quality of the ow,
X is calculated from energy balance, which requires
measurement of the mass ow rate, the temperature
of the liquid entering the heater, the heater power, and
the uid temperature in the test section. The HEM
void fraction
H
, is the simplest of the two-phase uid
modeling, whereby the gas and liquid phases are as-
sumed to be well mixed and velocity ratio S in Eq. (1)
is assumed to be unity. The average two-phase uid
density, is determined by:
=
G
+(1 )
L
(2)
The (HEM) uid density,
H
, is determined using
Eq. (2) by substituting
H
in place of . The (HEM)
pitch ow velocity, V
P
is determined by:
V
P
=
G
P

H
(3)
The pitch mass ux, G
P
, is determined from ow
measurements obtained from the orice plate reading
by:
G
P
=
( m/A
v
)P
(P D)
(4)
where m is the mass ow rate, A
v
is the upstream ow
area, P is the tube pitch, and D is the tube diameter.
238 S. Khushnood et al. / Nuclear Engineering and Design 230 (2004) 233251
Table 4
Representative published tests on two-phase ow
Researchers (published) Fluid Tube array Void
fraction
Tube
length
(mm)
Natural
frequency
(Hz)
Damping
ratio (%)
Pettigrew and Gorman
(1973)
Airwater Triangular/parallel,
square/rotated square
1020%
(quality) 50.8
17, 30 2.52.7
Heilker and Vincent
(1981)
Airwater Triangular/rotated
square
0.50.87
910
5662 0.84
Hara and Ohtani (1981) Airwater Single tube 0.020.61
60
Rigid
Remy (1982) Airwater Square 0.650.85 1000 56.6 0.61.75
Nakamura et al. (1982) Airwater Square/rotated square 0.20.94
190
142 1.31.7
Pettigrew et al. (1985) Airwater Triangular/square 0.050.98
600
2632 0.98.0
Axisa et al. (1984) Steamwater Square 0.520.98 1190 74 0.23.0
Nakamura et al. (1986) Steamwater Square 0.750.95
174
15.216 4.08.0
Hara (1987) Airwater Single/row 0.010.5
58
6.08.4 2.915.6
Goyder (1988) Airwater Triangular 0.50.8
360
175
Gay et al. (1988) Freon Triangular 0.580.84 1000 39.8 0.891.7
Nakamura and Fujita
(1988)
Airwater Square 0.020.95
600
52
2.1
Funakawa et al. (1989) Airwater Square/triangular 0.00.6
100
12.8
3.3
Nakamura et al. (1990) Steamwater Square 0.330.91
174
94137
Axisa et al. (1990) Steamwater Square/triangular
parallel/triangular
0.520.99 1190 72 0.24
Papp and Chen (1994) Normal
triangular/normal
square/parallel
triangular
2598%


Pettigrew et al. (1995) Freon Rotated triangular 4090%,
1090% 609
28150
0.15
Noghrehkar et al.
(1995)
Airwater Square fth and sixth
row
090%
200
1525 0.33.9
Taylor et al. (1995) Airwater U-bend tube bundle
with 180

U-tubes
parallel triangular
conguration
090% U-tube
radii
0.60.7
23114 1.52
Marn and Catton (1996) Airwater Normal triangular,
parallel triangular,
and rotated square
599%

0.721
Taylor and Pettigrew
(2000)
Freon Rotated triangular
and rotated square
5098%
609
0.25
Pettigrew et al. (2000) Airwater Normal 30

and
rotated 60

triangular,
normal 90

and
rotated 45

square
0100%
600
30160 15
Inada et al. (2000) Airwater Square 070% 198 15 2.5 1.6
Eq. added
damping
coefcient
S. Khushnood et al. / Nuclear Engineering and Design 230 (2004) 233251 239
Table 4 (Continued )
Researchers (published) Fluid Tube array Void
fraction
Tube
length
(mm)
Natural
frequency
(Hz)
Damping
ratio (%)
Nakamura et al. (2000)
(summary of Takai)
Freon 46 5 U-bend tubes,
specication of actual
Westinghouse type-51
series steam generator
(Not
considered)
based on
Connors
single-phase
relation

1626 Damping
ratio <1
Feenstra et al. (2000) R-11 Parallel/triangular 00.99

0100 1.12.9
Results 19731993 (Nakamura et al., 1993).
The main problem with using the (HEM) is that it
assumes zero velocity ratio between the gas and liquid
phases. This assumption is not valid in the case of
vertical upward ow, because of signicant buoyancy
effects.
4.2. Homogeneous ow (Taylor and Pettigrew, 2000)
This model assumes no relative velocity between
the liquid velocity U
l
and the gas velocity U
g
:
Slip = S = 1 : U
h
= U
g
= U
l
,
g
=
j
g
j
g
+j
l
(5)
where U
h
is the homogeneous velocity,
g
is the ho-
mogeneous void fraction, j
g
is supercial gas velocity
and j
l
is the supercial liquid velocity.
4.3. Smith correlation
Smith (1968) assume that kinetic energy of the liq-
uid is equivalent to that of the two-phase mixture and
a constant fraction K of liquid phase is entrained with
the gas phase. The value K = 0.4 was chosen to cor-
respond with the best agreement to experimental data
for ow in a vertical tube. Using the smith correlation,
the slip is dened as follows:
S = K +(1 K)
_
(X
l
/
g
) +K(1 X)
X+K(1 X)
_
1/2
(6)
where X is the mass quality,
g
is the density of the
gas phase and
l
is the density of the liquid phase.
4.4. Drift-ux model
The main formulation of drift-ux model was devel-
oped by Zuber and Findlay (1965). This model takes
into account both the two-phase ow non-uniformity
and local differences of velocity between the two
phases. The slip is dened as follows:
S =
(1
g
)
(1/(C
0
+(U
gj
/j))
g
)
=
(C
0
+(U
gj
/j)) (X/(X(1
g
/
l
))+(
g
/
l
))
(1 (X/(X(1
g
/
l
)) +(
g
/
l
)))
(7)
Out of the four unknown values of this relation,
two are determined experimentally: X directly and j
indirectly from mass ux measurement m:
j = j
g
+j
l
= m
_
X

g
+
(1 X)

l
_
(8)
The remaining two unknowns are empirical and
Lellouche and Zolotars Correlation (Lellouche and
Zollotar, 1982) is used to estimates these. This corre-
lation is well adapted for tube bundles in two-phase
cross-ow. An averaged gas phase drift velocity is de-
ned as follows:
U
gj
= 1.41
__

2
l
_

_
1/4
(1
g
)
1/2
(1 +
g
)
(9)
where is the liquid surface tension.
C
0
=
L
K
1
+(1 K
1
)
r
g
(10)
240 S. Khushnood et al. / Nuclear Engineering and Design 230 (2004) 233251
with
K
1
= K
0
+(1 K
0
)
_

l
_
1/4
,
K
0
= (1 +e
(Re/10
5
)
)
1
r =
_
1 +1.57
(
g
/
l
)
(1 K
0
)
_
, L =
(1 e
C
1

g
)
(1 e
C
1
)
,
C
1
=
4P
2
cr
P(P
cr
P)
where P
cr
is the two-phase ow critical pressure. Re =
UD/, where D is the tube diameter and is the mix-
ture viscosity which can be evaluated by McAdams
equation:
=

l
1 +
g
(
l
/
g
1)
(11)
where
l
and
g
are the viscosities of liquid and gas
phases, respectively.
4.5. Schrage correlation
The correlation by Schrage (1988) is based on em-
pirical data from an experimental test section, which
measures void fraction directly. This test section has
two valves capable of isolating a part of the ow al-
most instantaneously.
The correlation is based on physical considerations
and assumes two hypotheses:
1. For very low qualities, the gas phase appears only
as very small bubbles. The ow behaves as a homo-
geneous one and the minimum reduced void frac-
tion,
g
/
gh
value is 0.1.
2. For very high qualities, when X = 1, ow is con-
sidered homogeneous and we obtain the limit con-
dition
g
/
gh
= 1.
Then, the Schrage correlation is as follows:

gh
= 1 +0.123Fr
0.191
ln X with Fr =
m

gD
(12)
This correlation was established with an airwater
mixture but it remains valid for any other phase ow.
4.6. Feenstra model (Feenstra et al., 2000)
In this model predicted velocity ratio of the phases
is given by:
S = 1 +25.7(Ri Cap)
0.5
_
P
D
_
1
(13)
where the velocity ratio S is used in conjunction with
Eq. (1) for void fraction. The Richardson number Ri
is calculated by:
Ri =

2
ga
G
2
P
(14)
where a is the gap between tubes, ( =
l

g
)
is the density difference between the phases, g is the
gravitational acceleration and G
P
is the mass ow. The
capillary number (Cap) is calculated by:
Cap =

L
U
G

(15)
where
L
is the liquid phase absolute viscosity, is
the surface tension of liquid and U
G
is the gas phase
velocity.
4.7. Comparison of void fraction models
The (HEM) greatly over-predicts the actual gamma
densitometer void fraction measurement and the pre-
diction of void fraction model by Feenstra et al., is
superior to that of other models. It also agrees with
data in literature for airwater over a wide range of
mass ux and array geometry (Feenstra et al., 2000).
5. Dynamic parameters
5.1. Hydrodynamic mass
Hydrodynamic mass m
h
is dened as the equivalent
external mass of uid vibrating with the tube. It is
related to the tube natural frequency f in two-phase
mixture as discussed in (Carlucci and Brown, 1983)
and is given by:
m
h
= m
t
_
_
f
g
f
_
2
1
_
(16)
where m
t
is the mass of tube alone and f
g
is the natu-
ral frequency in air. The tube frequency is measured at
S. Khushnood et al. / Nuclear Engineering and Design 230 (2004) 233251 241
Fig. 1. Added mass as a function of void fraction (Carlucci, 1980).
mass uxes sufciently below uid-elastic instability.
Signicant shifts in frequency are observed at insta-
bility.
Hydrodynamic mass depends on the ratio P/d of the
tube pitch P to diameter d, given by Pettigrew et al.
(1989a):
m
h
=
_
d
2
4
__
(D
e
/d)
2
+1
(D
e
/d)
2
1
_
(17)
where is the two-phase mixture density.
D
e
d
=
_
0.96 +0.5P
d
_
P
d
, for a triangular bundle
D
e
d
=
_
1.07 +0.56P
d
_
P
d
, for a square bundle
where D
e
is equivalent diameter to model connement
due to the surrounding tubes as given by Rogers et al.
(1984).
Early airwater studies (Carlucci, 1980) showed
that added mass decreases with the void fraction as
shown in Fig. 1. It is also less than (1 ) where
is the void fraction. This deviation from expected
(1 ) line is caused by the air bubble concentrate
at the ow passage center. Added mass in two-phase
still receives little attention.
Fig. 2. Damping ratio as a function of void fraction (Carlucci and
Brown, 1983).
5.2. Damping
Subtracting the structural damping ratio from
the total yields the two-phase uid-damping ra-
tio (Noghrehkar et al., 1995). Total damping in-
cludes structural damping, viscous damping and
a two-phase component of damping as explained
by Pettigrew et al. (1994). The damping ratio in-
creases as the void fraction increases and peaks at
60% (Carlucci and Brown, 1983), then the ratio de-
crease with (Fig. 2). Damping also decrease as
the vibration frequency increase (Pettigrew et al.,
1985).
Damping in two-phase is very complicated. It is
highly dependent upon void fraction and ow regime.
The results for the two-phase component of damping
can be normalized to take into account the effect of
connement due to surrounding tubes by using the
connement factor C (Pettigrew et al., 2000). This
factor is a reasonable formulation of the connement
due to P/D. As expected, greater connement due to
smaller P/Ds increase damping. The connement fac-
tor is given by:
C =
1 +(D/D
e
)
3
[1 (D/D
e
)
2
]
2
(18)
242 S. Khushnood et al. / Nuclear Engineering and Design 230 (2004) 233251
6. Flow regimes
Many researchers have attempted the prediction of
owregimes in two-phase vertical ow. As yet, a much
smaller group has examined owregimes in cross-ow
over tube bundles. Some of the rst experiments were
carried out by Grant (Collier, 1979) as it was the only
available map at the time. Early studies in two-phase
cross-ow used the Grant map to assist in identify-
ing tube bundle ow regimes (Pettigrew et al., 1989b;
Taylor and Pettigrew, 2000). More recently (Ulbrich
and Mewes, 1994) performed a comprehensive anal-
ysis of available ow regime data resulting in ow
regime boundaries that cover a much larger range of
ow rates. They found that their new transition lines
had an 86% agreement with available data. Their ow
map is shown in Fig. 3 by Feenstra et al. (1996) with
the ow regime boundary transitions in solid lines and
the ow regimes identied with upper-case text. The
dotted lines outline a previous ow regime map based
on Freon-11 ow in a vertical tube, from (Taitel et al.,
1980).
Almost every study of ow regimes in tube bundles
has concluded that three distinct ow regimes exist.
In fact, several studies have shown that these regimes
can easily be identied by measuring the probability
density function (PDF) of the gas component of the
Fig. 3. Flow regime map for vertically upward two-phase ow from
(Feenstra et al., 1996, 1980). Data symbolize: square (Pettigrew
et al., 1989a,b), upward triangle (Axisa, 1985), downward triangle
(Pettigrew et al., 1995), and circle (Feenstra et al., 1995).
ow (Ulbrich et al., 1997; Noghrehkar et al., 1995;
Lian et al., 1997).
7. Tube to restraint interaction (wear work-rate)
Signicant tube-to-restraint interaction can lead to
fretting wear. Large amplitude out-of-plane motion
will result in large impact forces and in-plane motion
will contribute to rubbing action. Impact force and
tube-to-restraint relative motion can be combined to
determine work-rate. Work-rate is calculated using the
magnitude of the impact force and the effective slid-
ing distance during line contact between the tube and
restraint (Taylor et al., 1995). The work rate is given
in Eqs. (19) and (20):
W =
1
T
s
_
n
i=0
F
i
dS
i
(19)
W =
1
T
s
n

i=0
F
i
S
i
=
1
T
s
n

i=0
F
i
+F
i+1
2
S
i
(20)
where F is the instantaneous normal force, S is the
sliding distance during line contact and n is the num-
ber of points discretized over the sample duration T
s
.
As the work-rate increases, the effective wear rate in-
creases and the operational life of the U-bend tube
decreases. Implementation of the technology is de-
scribed in detail by Fisher et al. (1991). Measured val-
ues of wear work-rate for pitch velocity and mass ux
(Taylor et al., 1995) are presented in Fig. 4a and b,
respectively. The effect of uid-elastic forces is very
evident in the measured work-rates.
It is interesting to note that at higher pitch veloci-
ties and/or mass uxes, the wear work-rate does not
increase. Further study is required to understand why
the ow-rates do not affect the work-rates. This may
be related to the fact that at high void fractions and
high ow rates the random excitation forces are con-
stant with increasing ow rate (Taylor, 1992).
8. Fluid-elastic instability
Pettigrew and Gorman (1973) established the
fundamental treatment of two-phase ow induced
vibration. That was based on the method used for
S. Khushnood et al. / Nuclear Engineering and Design 230 (2004) 233251 243
Fig. 4. (a) Measured work-rate vs. pitch velocity (Taylor et al., 1995). (b) Measured work-rate vs. mass ux (Taylor et al., 1995).
single-phase ow-induced vibration. It was con-
cluded that uid-elastic instability seems possible in
two-phase ow as well as in liquid ow. The same in-
stability criterion seemed to apply to both liquid and
two-phase ow according to the following equation
in terms of dimensionless ow velocity U
p
/fD and
dimensionless mass-damping term [2m/D
2
]
1/2
.
U
p
fD
= K
_
m
D
2
_
1/2
(21)
where K is the instability factor, U
p
the average ow
velocity, the average uid density, the average
damping in two-phase ow and f is the tube natural
frequency in two-phase ow mixture at mass ux m
p
,
with u
p
= m
p
/ and is the total damping ratio.
Fluid-elastic instability is possible when the uid
dynamic forces on the tubes are proportional to the
tube motion. When energy absorbed form the uid by
the tubes exceeds the energy dissipated by damping,
uid-elastic instability results in very large vibration
amplitudes. A complete, stability boundary estimation
method has not yet been established, at least not in for-
mal theoretical manner (Nakamura et al., 2000). A de-
sign procedure has already been proposed by Connors
(1970) and it is widely believed to be useful for de-
signers, although there have been some modications
of his parameters such as the critical factor K, and the
exponents in Eq. (21).
The distribution of density and uid ow velocity
is estimated using the following equation proposed by
Connors (1978):
U
c
f
i
D
= K
_
m
0

0
D
2
S
i
_
1/2
(22)
where
S
i
=
_
L
0
((S)/
0
)(S)
2

i
(S)
2
ds
_
L
0
(m(S)/m
0
)
i
(S)
2
ds
where (S) is the ow distribution,
i
(S) is the corre-
sponding shape function, L is the limit span subjected
to cross-ow and i refers to mode number.
Eq. (22) is theoretically derived from the hypoth-
esis that the uid-elastic force has a linear relation
to the tube displacement; this implies the so called
displacement mechanism (Chen, 1987). However, it
was late established in the 1980s that the mechanism
of uid-elastic instability is not limited to displace-
ment mechanism but also that a velocity mechanism
exists. Eq. (22) therefore cannot be applied to every
244 S. Khushnood et al. / Nuclear Engineering and Design 230 (2004) 233251
case. This is true for single-phase ow. Pettigrew et al.
(1978) considered the effect of two-phase ow, but
Connors derived the analysis from the same equation.
Some papers have been published for the estimation
of the critical factor K in two-phase ow condition
based on Eq. (21) by Antunes et al. (1985). Nakamura
et al. (1997, 2000) have tried to develop a new method
with different approaches to this problem. This method
is based on the insight into the physical aspect of
the two-phase phenomenon. The method is essentially
macroscopic in nature being based on Connors dis-
placement mechanism.
Eq. (23) gives the stability boundary for the case of
homogeneous two-phase ow.

2
eq
+
2
eq
= 1 + (23)
where parameter = (m
geq
/m
leq
) is dened to be
the ratio of non-dimensional effective masses per unit
length of tube in liquid phase and gas phase, respec-
tively. Eq. (23) is the energy balance in simple form.

eq
and
eq
are the non-dimensional effective ow ve-
locities of the liquid phase and gas phase, respectively,
where = ( /1 ) (non-dimensional parameter),
(s). T is the time length of gas phase, the aver-
age value of (s), a non-dimensional parameter and
T is the time period. = (C

g
/C

l
)(
g
/
l
)(U
cg
/U
cl
)
2
,
where C

l
, C

g
are non-dimensional uid force coef-
cients, U
cl
and U
cg
are critical ow velocities for dif-
ferent phases.
Fig. 5. Vibration response at 80% void fraction: exible vs. rigid tube bundle (Pettigrew et al., 1995).
Fig. 6. Instability results in two-phase cross-ow: comparison of
Freon vs. airwater (Pettigrew et al., 1995).
In two-phase Freon, the critical mass ux for insta-
bility is reasonably well dened as shown in Fig. 5.
Instability results (Pettigrew et al., 1995) are presented
in Fig. 6. Generally, the results show two regions of in-
stability. The rst is a region where the mass-damping
parameter exponent is roughly 0.5 as expected. These
regions occur at void fractions below approximately
80% for airwater mixtures, and below approximately
65% for Freon two-phase ow. For this region, the in-
stability factor K is above the recommended guideline
K = 3.0. There appears another instability region at
void fractions above 80% for airwater and above
65% for Freon, where the slope of instability curve
S. Khushnood et al. / Nuclear Engineering and Design 230 (2004) 233251 245
is considerably lower. In fact, the change in behavior
is dramatic for Freon two-phase ow. At 90% void
fractions the instability factor K = 1.36, which is less
than the recommended design guidelines of K = 3.0.
9. Random turbulence excitation
For a given void fraction, the vibration response
of the tube bundle generally increase with mass ux
until uid-elastic instability is reached as shown in
Fig. 7ad, which gives typical comparative (exible
versus rigid tube bundle) vibration response (RMS)
at the tube free end versus mass ux. The differ-
ence in vibration response between exible bundle and
Fig. 7. (ad) Typical RMS vibration response: comparison between exible and rigid tube bundle (Pettigrew et al., 2000).
rigid bundle is small at mass uxes below instabil-
ity, which means that random turbulence excitation is
not much effected by surrounding tubes. On the other
uid-elastic instability occurs at lower mass uxes for
exible tube bundle (Pettigrewet al., 2000). The vibra-
tion response below instability is attributed to random
turbulence excitation. The approach is to deduce the
randomturbulence excitation forces fromthe vibration
response (Pettigrew et al., 2000). The relationship be-
tween the mean square amplitude of the tube vibration
response (Y
2
) and the power spectral density S
F
(f) of
the random excitation forces can be developed from
random vibration theory. For the fundamental mode of
a cantilevered tube subjected to uniformly distributed
and spatially correlated forces over its entire length,
246 S. Khushnood et al. / Nuclear Engineering and Design 230 (2004) 233251
Y
2
(l) by Mirza and Gorman (1973) is given by:
Y
2
(l) = 0.613S
F
(f)/(16
2
f
3
m
2
) (24)
where m is mass per unit length including that of tube
m
t
and hydraulic mass m
h
, is modal damping ra-
tio, f is natural frequency of tube, and C = 0.613 is
the coefcient of uid tube interaction which varies
with the tube boundary conditions and axial location x.
For Taylors experimental conditions (Cantilever tube)
maximum displacement occur at x = l, where l is
the tube length. Thus, it is easy to deduce S
F
(f) from
the vibration response with Eq. (24). A relationship
(Inada et al., 2000) between the vibration response
and the mass ux m
p
was established by doing a lin-
ear regression analysis on the experimental data. By
choosing only the data points corresponding to mass
uxes below the threshold for uid-elastic instability
a mass ux exponent b was determined for the rela-
tion (Y
2
)
0.5
m
b
p
. The scatter in the results was sig-
nicant, but a tendency towards a value of b = l was
found in the void fraction range of 2590%. At void
fraction below 25% the plots were extremely difcult
to interpret, due to the presence of some periodic wake
shedding peaks, as discussed in Taylor et al. (1989).
Pettigrew et al. (2000) uses an exponent of b = l to
collapse the random turbulence data. This allowed for
the calculation of a simplied normalized power spec-
tral density NPSD of the excitations forces formulated
Fig. 8. (a and b) Normalized power spectral density of random turbulence excitation (Pettigrew et al., 2000).
by:
NPSD =
S
F
(f)
( m
p
D)
2
(25)
The normalized power spectral densities (Nakamura
et al., 1982) were plotted against void fraction in
Fig. 8a and b. A comparison between P/D = 1.22
and 1.47 indicates that random turbulence forces are
not greatly affected by P/D. This is not entirely un-
expected since the response is not much affected by
the motion of surrounding tubes as shown in Fig. 7,
which gives typical RMS vibration response. Fig. 8a
and b illustrates that the random turbulence levels are
practically the same for triangular and square bundles
for the worst orientation. The curve (Taylor et al.,
1989) is dened by:
NPSD = 10
(0.03
g
5)
, for 25 <
g
< 90% (26)
Total owpressure is a sumof dynamic pressure p
D
,
and static pressure p
S
. Flow-induced vibration must
dissipate energy from this ow. Therefore, tube vi-
bration must require a decrease in total ow pressure.
When other owparameters remain the same, the mea-
sure of dissipated energy is a decrease in static pres-
sure. Papp (1988) proposed the following relationship:
(Y
2
)
0.5
pd (27)
where p is the pressure drop across a tube row, and
d is the tube diameter. Of course, there is energy dissi-
S. Khushnood et al. / Nuclear Engineering and Design 230 (2004) 233251 247
pation in the form of heat; but it is not expected to be
important to tube vibration. From Eqs. (24) and (27)
the following relation can be written:
S
F
(f) = C
2
I

2
p d
2
(28)
The two-phase owpressure drop is given by Hewitt
(1990) (Section 2.3.2):
p
TPF
=
2
L0
Eu
L0
Q
2
2
L
(29)
where
L
is liquid density,
2
L0
is the two-phase fric-
tion multiplier and Eu
L0
the Euler number, is related
only to the pressure drop of liquid ow (total mass ux
represented by liquid) by p
L0
= Q
2
/(2
L
). Substi-
tuting Eqs. (28) and (29) in (24) gives:
(Y
2
(x))
0.5
= C
I

2
L0
Eu
L0
_
C
16f
3
m
2

_
1/2
Q
2
2
L
d
(30)
where C
I
is coefcient of interaction.
It is evident that Eq. (30) takes into account all the
parameters of a coupled uid/tube system, i.e.
(a)
L0
is a function of two-phase ow parameters
(void fraction , mass ux Q, and pressure p);
(b) Eu
L0
depends on tube bundle geometry; and
(c) The parameters within the parentheses on the
right-hand side of the equation are determined by
the properties of the tube and its end condition.
The main source of uncertainty in Eq. (30) is the
lack of data on the two-phase friction multiplier
2
L0
(Wambsganss et al., 1992). For a tube bundle (Hewitt,
1990), (Section 2.3.2) recommended the Chisholm
correlation given by:

2
L0
=1 +(Y
2
1)
[BX
(2n)/2
(1 X)
(2n)/2
+X
(2n)
] (31)
Table 5
Values of the coefcient of interaction C
I
(Papp and Chen, 1994)
(%) Normal triangular
array P/d = 1.47
Normal triangular
array P/d = 1.32
Normal square
array P/d = 1.47
Parallel triangular
array P/d = 1.47
25 0.0260 0.0193 0.0379 0.0280
50 0.0278 0.0149 0.0326 0.0340
75 0.0458 0.0353 0.0591 0.0607
90 0.0460 0.0308 0.0599 0.0725
98 0.0467 0.0334
a
0.0716 0.0591
where X is ow quality, Y
2
= p
G0
/p
L0
=
Eu
G0L
/Eu
L0
G
0
, (p
G0
) is the pressure drop for
the gas owing alone, n = 0.37 and B = 1. Nev-
ertheless, intermediate ow regimes are typical for
the experimental data used. Therefore, the original
Chisholm recommendation was used for
2
L0
calcula-
tions. According to Chisholms recommendation, the
following relationships were used for parameter B:
B =
520
YQ
1/2
for Q 600 kg/m
2
s (32a)
and
B =
21
Y
for Q > 600 kg/m
2
s (32b)
Using recommended equations for Euler number
(Hewitt, 1990, Section 2.2.4), the values of n for
Re = 100, 1000 and 10,000 were calculated. In this
range of Reynolds number, the value of n changes
from n = 0.86 (Re = 100) to 0 (Re = 10,000; nor-
mal square conguration). Therefore, the interval of
Reynolds number was divided into three particular
intervals with the following values of n:
Re 650 (Q 50 kg/m
2
s), n = 0.8
650 < Re < 2000
(50 < Q 154 kg/m
2
s), n = 0.5
Re 2000 (Q 154 kg/m
2
s), n = 0.25
(33)
Using these values of n and Eqs. (31) and (32), the
two-phase pressure drop multiplier was calculated. Fi-
nally from Eq. (30) the values of C
I
were calculated.
These values (average for each geometry and void
fraction) are given in Table 5. It is evident that for each
geometry we obtained two intervals of void fraction,
= 2550 and 7598%. In each interval, the value of
C
I
can be regarded as constant; i.e. the condition of
validity of correlation is validated.
248 S. Khushnood et al. / Nuclear Engineering and Design 230 (2004) 233251
10. Measurement of void fraction
In general, the surveyed research indicates two
types of void fraction measurements (Feenstra et al.,
2000), the HEM void fraction and Radiation Atten-
uation Method (RAD) void fraction. HEM refers to
Homogeneous Equilibrium Model and RAD refers
to Radiation Attenuation Method. The determination
of uid parameters (uid density and ow velocity)
are quite different when these two methods are used
(Feenstra et al., 2000). In RAD method (Feenstra
et al., 2000; Chan and Bannergee, 1981) gamma ux
from radiation source which penetrates the test section
will be attenuated by different amounts depending
upon the average density of the two-phase ow. Void
fraction can be determined by interpolating the
average density of the uid between the benchmark
measurements for 100% liquid and gas according to
the following equation:
=
ln(N/N
L
)
ln(N
G
/N
L
)
(34)
where N represents the gamma counts obtained dur-
ing an experimental trial, N
L
and N
G
are the reference
counts obtained prior to the experiment for 100% liq-
uid and 100% gas, respectively. U
G
and U
L
can be
Table 6
Results of example FIV analysis (TEMA, 8th ed.)
calculated by:
U
G
=
xG
P

G
(35)
U
L
=
(1 x)G
P
(1 )
L
(36)
Alogical measure of an equivalent two-phase veloc-
ity, V
eq
determined from averaging the dynamic head
of the gas and liquid phases is given by Eq. (37):
V
eq
=
_

G
U
2
G
+(1 )
L
U
2
L

(37)
Local void fraction measurements have been pub-
lished by Haquet and Gouirand (1995), Lian et al.
(1997), and Mann and Mayinger (1995).
11. Application of TEMA (FIV) code
Table 6 give the results summary of application of
TEMA (2003) (FIV) software to a U-tube span sub-
jected to two-phase cross-ow. Summary of input ge-
ometric and process parameters is given. Analysis re-
sults for various mechanisms, logarithmic decrement
and acceptable amplitudes are also presented.
S. Khushnood et al. / Nuclear Engineering and Design 230 (2004) 233251 249
12. Conclusions
The current paper has reviewed the two-phase
cross-ow induced vibration in tube bundles. Follow-
ing topics have been considered.
Types of two-phase ow and ow regimes.
Choice of modeling uids, their comparison from
experimental standpoint and dynamic parameters
such as damping and added mass.
Various void factors models, their comparison and
void fraction measurement.
Tube vibration excitation mechanisms induced by
uid-elastic instability and random turbulence.
Wear work rates for tube-to-restraint interaction.
There is a strong need for establishing reliable
design procedures for two-phase cross-ow tube
bundle vibrations. This could be achieved by carry-
ing out modeling and simulation of the system with
uidstructure interaction focusing on void fraction,
and reliable experimental data. Test data on high
pressure and temperature conditions are insufcient,
therefore a potential challenge lies ahead.
Acknowledgements
The authors are indebted to College of Electrical &
Mechanical Engineering, National University of Sci-
ences & Technology (NUST), Pakistan, and Pakistan
Science Foundation (PSF) for the completion of this
research work. We also gratefully acknowledge the
help provided by Pakistan Scientic & Technical In-
formation Center (PASTIC) in acquiring the related
literature.
References
Antunes, J., Villard, B., Axisa, F., 1985. Cross-ow induced
vibration of U-bend tubes of steam-generator. Proc. SMiRT
F-16/9, 229239.
Axisa, F., 1985. Vibration of tube bundles subjected to steamwater
cross-ow: a comparative study of square and triangular
arrays. In: Paper B1/2, 8th International Conference on SMiRT,
Brussels.
Axisa, F., Villard, B., Gibert, R.J., Hostroni, G., Sundheimer,
P., 1984. Vibration of tube bundles subjected to airwater
and steamwater cross-ow. In: Proceedings of ASME Winter
Annual Meeting, p. 269.
Axisa, F., Antunes, J., Villard, B., 1990. Random excitation of
heat exchanger tubes by cross-ows. J. Fluid Struct. 4, 321.
Carlucci, L.N., 1980. Damping and hydrodynamic mass of a
cylinder in simulated two-phase ow. J. Mech. Des. 102, 597
602.
Carlucci, L.N., Brown, J.D., 1983. Experimental studies of
damping and hydrodynamic mass of a cylinder in conned
two-phase ow. ASME J. Vibrat. Acoust. Stress Reliability Des.
105, 8389.
Chan, A.M.C., Bannergee, S., 1981. Design aspects of gamma
densitometers for void fraction measurements in small scale
two-phase ow. Nuclear Instum. Methods 190, 135148.
Chen, S.S., 1987. A general theory for dynamic instability of tube
arrays in cross-ow. J. Fluids Struct. 1, 3553.
Collier, J., 1979. Convective Boiling and Condensation. Oxford
Science Publication.
Connors, H.J., 1970. Fluid-elastic vibration of tube arrays excited
by cross-ow. In: Proceedings of ASME Winter Meeting,
pp. 4256.
Connors, H.J., 1978. Fluid-elastic vibration of heat exchanger tube
arrays. ASME J. Mech. Des. 100, 347353.
Feenstra, P.A., Judd, R.L., Weaver, D.S., 1995. Fluid-elastic
instability in a tube array subjected to two-phase R-11
Cross-ow. J. Fluids Struct. 9, 747771.
Feenstra, P.A., Weaver, D.S., Judd, R.L., 1996. Stability analysis of
parallel triangular tube arrays subjected to two-phase cross-ow.
In: Proceedings of CSME Conference, Manchester University.
Feenstra, P., Weaver, D.S., Judd, R.L., 2000. Modeling two-phase
ow excited uid-elastic instability in tube arrays. In:
Proceedings of the Seventh International Conference on
Flow-Induced Vibrations, FIV-2000, Lucerne, Switzerland, June
1922, ISBN 90 5809 1295, pp. 545554.
Fisher, N.J., Ing, J.G., Pettigrew, M.J., Rogers, R.J., 1991. Fretting
wear damage prediction in the inlet region of nuclear steam
generators. In: Proceedings of I. Mech E. Conference, Paper C
416/053, London.
Frick, T.M., Sobek, T.E., Reavis, J.R., 1984. Overview on the
development and implementation of methodologies to compute
vibration and wear of steam generator tubes. In: Proceedings
of the ASME Symposium on Flow-Induced Vibrations, vol. 3,
New Orleans.
Funakawa, M., Ishimatsu, T., Kumon, K., Edakuni, K., 1989.
Excitation mechanism of uid-elastic vibration in tube array.
Trans. Jpn. Soc. Mech. Eng. 55 (519C), 2718 (in Japanese).
Gay, N., Decembre, P., Launay, J., 1988. Comparison of airwater
to water-freon two-phase cross-ow effects on the vibrationary
behaviour of a tube bundle. In: Proceedings of ASME Winter
Annual Meeting, p. 139.
Goyder, H.G.D., 1988. Fluid-elastic instability and buffeting of
heat exchanger tube bundles due to single and two-phase ows.
In: Proceedings of ASME Winter Annual Meeting, p. 151.
Haquet, J.F., Gouirand, J.M., 1995. Local two-phase ow
measurements in a cross-ow steam-generator tube bundle
geometry. In: The Minnie II XF Program, Advances in
Multiphase Flow. Elsevier Science B.V.
Hara, F., 1987. Vibration of a single row of circular cylinders
subjected to two-phase bubble cross-ow. In: Proceedings of
250 S. Khushnood et al. / Nuclear Engineering and Design 230 (2004) 233251
the International Conference of Flow Induced Vibration, No.
E1, p. 203.
Hara, F., Ohtani, I., 1981. Study on vibration of a cylinder in
two-phase ow. Trans. Jpn. Soc. Mech. Eng. 48 (431C), 962
(in Japanese).
Heilker, W.J., Vincent, R.Q., 1981. Vibration in nuclear heat
exchangers due to liquid and two-phase ow. J. Eng. Power
ASME 103, 358.
Hewitt, G.F., 1990. Handbook of Heat Exchanger Design.
Hemisphere publishing Co., New York.
Inada, F., Kawamura, K., Yasuo, A., Yoneda, K., 2000. An
experimental study on the uid-elastic forces acting on a square
tube bundle in two-phase cross-ow: the effect of vibration
amplitude. In: Proceedings of FIV-2000, Lucerne, Switzerland,
June 1921, ISBN 90 5809 1295, pp. 555568.
Lellouche, G., Zollotar, B., 1982. Mechanistic model for predicting
two-phase void fraction for water in vertical tubes, channels, and
rod bundles. Electric Power Research Institute Report NP-2246.
SR.
Lian, H.Y., Noghrehker, G., Chan, A.M.C., Kawaji, M., 1997.
Effect of void fraction on vibrational behaviour of tubes in
tube bundle under two-phase cross-ow. J. Vibrat. Acoust. 119,
457463.
Mann, W.F., Mayinger, F., 1995. Flow induced vibration of
tube bundles subjected to single and two-phase cross-ow. In:
Advances in Multiphase ow. Elsevier Science B.V.
Marn, J., Catton, I., 1996. On analysis of uid-elastic instability
in airwater mixture. J. Pres. Ves. Technol. 118, 188193.
Mirza, S., Gorman, D.G., 1973. Experimental and analytical
correlation of local driving forces and tube response in liquid
ow induced vibration of heat exchangers. In: Proceedings of
the Second SMiRT Conference, Paper No. F615.
Nakamura, T., Fujita, K., 1988. Study of ow-induced vibration
of a tube array by a two-phase ow (3rd report). Trans. Jpn.
Soc. Mech. Eng. 54 (502C) 1204 (in Japanese).
Nakamura, T., Fujita, K., Shiraki, K., Kanazawa, H., Sakata, K.,
1982. An experimental study on exciting force by two-phase
cross-ow. Proc. Pres. Ves. Piping Conf. ASME 63, 19.
Nakamura, T., Fujita, K., Kawanishi, K., Saito, I., 1986. A study
of the ow-induced vibration of a tube array by a two-phase
ow (2nd report). Trans. Jpn. Soc. Mech. Eng. 52 (483C), 2790
(in Japanese).
Nakamura, T., Kawanishi, K., Fujita, K., 1990. Study on
ow-induced vibration of a tube array by a two-phase ow
(4th report). Trans. Jps. Soc. Mech. Eng. 56 (525C) 1164 (in
Japanese).
Nakamura, T., Fujita, K., Tsuge, A., 1993. Two-phase cross-ow
induced vibration of tube arrays. JSME Int. J. Ser. B 36 (3),
429438.
Nakamura, T., Murerithi, N.W., Chen, S.S., 1997. Fluid-elastic
instability threshold for tube arrays in two-phase ow: a model
incorporating temporal uid force correction along the tube
axis. In: Proceeding of ASME Winter Annual Meeting, A.D
vol. 5362, pp. 365372.
Nakamura, T., Hirota, K., Tomomatsu, K., 2000. Some problems
on the estimation of ow-induced vibration of a tube array
subjected to two-phase ow. In: Proceedings of FIV-2000, ISBN
90 5809 1295, pp. 537544.
Noghrehkar, R., Kawaji, M., Chan, A.M.C., 1995. An experimental
study of local two-phase parameters in cross-ow induced
vibration in tube bundles. In: Proceedings of the FIV-95, Sixth
International Conference on Flow Induced Vibrations, London,
April 1012, ISBN 90 5410 547x, pp. 373382.
Paidoussis, M.P., 1982. A review of ow induced vibrations in
reactors and reactor components. Nuclear Eng. Des. 74, 3640.
Papp, L., 1988. Vibration of VVER NPP steam generator tubes.
Jadera Energie 34 (7), 249259 (in Czech).
Papp, L., Chen, S.S., 1994. Turbulence-induced vibration of tube
arrays in two-phase ow. J. Pres. Ves. Technol. 116, 312316.
Pettigrew, M.J., Gorman, D.J., 1973. Experimental studies on
ow-induced vibration to support steam generator design. Part
III. In: Proceedings of the International Symposium on Vib.
Prob. in Industry, AECL-5804.
Pettigrew, M.J., Taylor, C.E., 1994. Two-phase ow induced
vibrations: an overview (survey paper). ASME J. Pres. Ves.
Technol. 116, 233253.
Pettigrew, M.J., Sylvestre, Y., Campagna, A.O., 1978. Vibration
analysis of heat exchanger and steam generator designs. Nuclear
Eng. Des. 48, 99115.
Pettigrew, M.J., Tromp, J.H., Mastorakos, J., 1985. Vibration of
tube bundles subjected to two-phase cross-ow. J. Pres. Ves.
Technol. ASME 107, 335.
Pettigrew, M.J., Taylor, C.E., Kim, B.S., 1989a. Vibration of tube
bundles in two-phase cross-ow. Part 1. Hydrodynamic mass
and damping. ASME J. Pres. Ves. Technol. 111, 466477.
Pettigrew, M.J., Tromp, J.H., Taylor, C.E., Kim, B.S., 1989b.
Vibration of tube bundles in two-phase cross-ow. Part 2.
Fluid-elastic instability. J. Pres. Ves. Technol. 111, 478487.
Pettigrew, M.J., Carlucci, L.N., Taylor, C.E., Fisher, N.J., 1991.
Flow induced vibration and related technologies in nuclear
components. Nuclear Eng. Des. 131, 81100.
Pettigrew, M.J., Taylor, C.E., Yasuo, A., 1994. Vibration damping
of heat exchanger tube bundles in two-phase ow. Welding
Research Council, Bulletin No. 389, New York, NY, pp. 141.
Pettigrew, M.J., Taylor, C.E., Jong, J.H., Currie, I.G., 1995.
Vibration of tube bundles in two-phase Freon cross-ow. J.
Pres. Ves. Technol. 117, 321329.
Pettigrew, M.J., Taylor, C.E., Kim, B.S., 2000. The effects of
tube bundle geometry on vibration in two-phase cross-ow. In:
Proceedings of the International Conference on Flow-Induced
Vibrations, FIV-2000, Lucerne, Switzerland, June 1922, ISBN
90 5809 1295, pp. 561568.
Price, S.J., 1995. A review of theoretical models for uid-elastic
instability of cylinder arrays in cross-ow. J. Fluids Struct. 9,
463518.
Remy, F.N., 1982. Flow induced vibration of tube bundles in
two-phase cross-ow. In: Proceedings of the Third Keswick
International Conference on Vibration in Nuclear Plant. Paper
1.9, p. 135.
Rogers, R.G., Taylor, C., Pettigrew, M.J., 1984. Fluid effects
on multi-span heat exchanger tube vibration. In: ASME PVP
Conference, San Antonio, TX, ASME Pbl. H00316, Topics in
Fluid Structure Interactions, pp. 1726.
Schrage, D., 1988. Two-phase pressure drop in vertical cross-ow
across horizontal tube bundle. AIChE J. 34 (1), 107115.
S. Khushnood et al. / Nuclear Engineering and Design 230 (2004) 233251 251
Smith, S., 1968. Void fractions in two-phase ow: a correlation
based upon an equal velocity head model. Proc. Inst. Mech.
Eng.
Standards of Tubular Exchanger Manufacturers Association Inc.
(TEMA) Sec. 6, 8th ed. Flow Induced Vibration Program (FIV).
Taitel, Y., Barnea, D., Dukler, A., 1980. Modeling ow pattern
transition for steady upward gas-liquid ow in vertical tubes.
AIChE J. 26, 345354.
Taylor, C.E., 1992. Random excitation forces in tube arrays
subjected to two-phase cross-ow. In: Proceedings of
WAM-1992, International Symposium on Flow Induced
Vibration & Noise, vol. 1, HTD-vol. 230, pp. 89107.
Taylor, C.E., Pettigrew, M.J., 2000. Effect of ow-regime and
void fraction on tube bundles vibration. In: Proceedings of
FIV-2000, Lucrene, Switzerland, June 1921, ISBN 90 5809
1295, pp. 529536.
Taylor, C.E., Currie, I.G., Pettigrew, M.J., Kim, B.S., 1989.
Vibration of tube bundles in two-phase cross-ow. Part 3.
Turbulent induced excitation. ASME J. Pres. Ves. Technol. 111,
488500.
Taylor, C.E., Boucher, K.M., Yetisir, M., 1995. Vibrational impact
forces due to two-phase cross-ow in U-bend of nuclear
steam generators. In: Proceedings of FIV-95, Sixth International
Conference on Flow Induced Vibration, ISBN 90 5410 547x,
pp. 401411.
Ulbrich, R., Mewes, D., 1994. Vertical, upward gas liquid
two-phase ow across a tube bundle. Int. J. Multiphase Flow
20, 249272.
Ulbrich, R., Reinecke, N., Mewes, D., 1997. Recognition of ow
pattern for two-phase ow across tube bundle. In: Proceedings
of the Fourth World Conference on Experimental Heat Transfer,
Fluid Mechanics and Thermodynamics, Brussels, June 26.
Wambsganss, M.W., Jendrzejczyk, J.A., France, D.M., 1992.
Two-phase ow and pressure drop in ow passages of compact
heat exchangers. SAE Technical Paper Series 920550.
Weaver, D.S., FitzPatrick, J.A., 1988. A review of cross-ow
induced vibrations in heat exchanger tube arrays. J. Fluids
Struct. 2, 7393.
Zuber, N., Findlay, J., 1965. Average volumetric concentration in
two-phase ow systems. ASME J. Heat Transfer 87, 453468.