Y. A.

Khulief
1
e-mail: khulief@kfupm.edu.sa
S. A. Al-Kaabi
e-mail: kaabis@kfupm.edu.sa
S. A. Said
e-mail: samsaid@kfupm.edu.sa
M. Anis
e-mail: anisqasm@kfupm.edu.sa
Department of Mechanical Engineering,
King Fahd University of Petroleum & Minerals,
KFUPM Box 1767,
Dhahran 31261, Saudi Arabia
Prediction of Flow-Induced
Vibrations in Tubular Heat
ExchangersPart I: Numerical
Modeling
Flow-induced vibrations due to crossow in the shell side of heat exchangers pose a
problem of major interest to researchers and practicing engineers. Tube array vibrations
may lead to tube failure due to fretting wear and fatigue. Such failures have resulted in
numerous plant shutdowns, which are often very costly. The need for accurate prediction
of vibration and wear of heat exchangers in service has placed greater emphasis on the
improved modeling of the associated phenomenon of ow-induced vibrations. In this
study, the elastodynamic model of the tube array is modeled using the nite element
approach, wherein each tube is modeled by a set of nite tube elements. The interaction
between tubes in the bundle is represented by uidelastic coupling forces, which are
dened in terms of the multidegree-of-freedom elastodynamic behavior of each tube in
the bundle. Explicit expressions of the nite element coefcient matrices are derived. The
model admits experimentally identied uidelastic force coefcients to establish the nal
form of equations of motion. The nonlinear complex eigenvalue problem is formulated
and solved to determine the onset of uidelastic instability for a given set of operating
parameters. DOI: 10.1115/1.3006950
Keywords: ow-induced vibrations, heat exchangers, tube bundles, crossow
1 Introduction
It is well known that bundles of tubes used in heat exchangers
and boilers in nuclear and chemical plants can be excited to vi-
brate excessively when exposed to crossow. Pressure uctuations
around heat exchanger tubes result in uid-structure dynamic cou-
pling, which give rise to vibrations in each tube of the bundle.
Accordingly, the uid dynamic forces on one of the tubes in the
bundle are induced by the vibration of the tube itself and by the
vibration of the neighboring tubes. This is due to the fact that
pressure uctuations around the tubes result in uid-structure dy-
namic coupling. At sufciently high ow velocities, a large am-
plitude vibration at the natural frequencies of the tubes may take
place. This phenomenon is known as uidelastic instability and is
of great concern to plant operators. The onset of large amplitude
vibration is quite abrupt when the crossow velocity is increased
above the so-called critical velocity. These vibrations may ap-
proach resonance, thus leading to failure of the tubes. Damage
also occurs when the tube vibration results in midspan collisions
between adjacent tubes, as well as collisions between tubes and
bafe holes.
To gain more insight into this phenomenon, researchers resorted
to dynamic modeling of such systems. However, modeling the
dynamics of uidelastic motions due to crossow over a tube
bundle is too complicated to be investigated only analytically. At
present there is no such a reliable analytical model that accurately
describes the phenomena of ow-induced vibration over a bank of
tubes. Accordingly, a reliable mathematical model requires some
tuning via experimental measurements.
The ow-induced vibration phenomena of tube bundle vibra-
tions caused by shell-side ow in heat exchangers were addressed
early in the literature 1,2. Today, the literature in this area of
research has become very rich, with a large number of publica-
tions addressing these problems and suggesting different methods
of predictions and solutions. These include approximate analytical
models, analytical models with purely structural emphasis, semi-
analytical models where modeling simplications were attributed
to complementary experimental investigations, and pure experi-
mental studies dedicated to the understanding and identication of
such excitation mechanisms. Progress in research activities related
to this problem was reviewed by Paidoussis 3, Price 4, and
Weaver et al. 5. In addition, some books were dedicated to pre-
senting a rather detailed account of this problem, e.g., the books
by Chen 6 and Katinas and Zukauskas 7.
A number of theoretical investigations have been conducted on
ow-induced vibrations in heat exchangers. Chen 8,9, in his
sequel papers, presented a general theoretical approach to charac-
terizing the instability mechanisms of a group of tubes in cross-
ow. It was concluded that no single stability criterion could be
applicable to all cases for all ranges of parameters. Instead, dif-
ferent stability criteria have to be devised for different instability
mechanisms and different parameter ranges.
Cai et al. 10 reported a theoretical investigation of the u-
idelastic instability of loosely supported tubes in nonuniform
crossow, wherein the unsteady ow theory was employed. Their
instability analysis, however, was restricted to the inactive phase
of the tube motion. Further investigations were suggested to study
the instability mechanism during the active mode of the tube mo-
tion. Cai and Chen 11 extended the theoretical study of Cai et al.
10 to include chaotic vibrations associated with the uidelastic
instability of nonlinearly supported tubes in crossow. However,
they indicated that chaotic analysis may differ based on the com-
plexity of the model, and that results derived for a one-or two-
degree-of-freedom systems may not be true for multidegree-of-
freedom systems. Accordingly, further investigation of the
phenomenon was suggested. Eisinger et al. 12 applied a numeri-
1
Corresponding author.
Contributed by the Pressure Vessel and Piping Division of ASME for publication
in the JOURNAL OF PRESSURE VESSEL TECHNOLOGY. Manuscript received January 4,
2007; nal manuscript received April 27, 2008; published online November 10,
2008. Review conducted by David Raj.
Journal of Pressure Vessel Technology FEBRUARY 2009, Vol. 131 / 011301-1 Copyright 2009 by ASME
Downloaded 28 Nov 2008 to 212.26.1.29. Redistribution subject to ASME license or copyright; see http://www.asme.org/terms/Terms_Use.cfm
cal model to simulate the uidelastic vibration of a representative
tube in a tube bundle based on the unsteady ow theory. The
numerical simulation was performed using the general purpose
ABAQUS-EPGEN nite element code using a special subroutine in-
corporating uidelastic forces. The results were validated against
some published data based on linear cases.
A generalization of the quasisteady theory was developed by
Granger and Paidoussis 13 and was applied to study the cross-
ow induced vibration of tube arrays. This study was conducted
on the basis of the continuity and NavierStokes equations, in the
case of an impulsive movement of a body subjected to crossow.
Numerical solutions showed that the quasi-unsteady model was a
clear improvement on the conventional quasisteady approach,
leading to a more reasonable agreement with experimental results
and providing rened insights into the physical mechanisms re-
sponsible for uidelastic instability.
An analytical computational uid dynamics technique was in-
troduced by Ichioka et al. 14 to study the uidelastic vibration of
tube bundles in heat exchangers. The technique was based on the
moving mesh method, which was developed by the authors. The
technique was examined for the case of two cylinders in an in-
nitely arrayed cylinder row.
Kassera and Strohmeier 15 introduced a two-dimensional
simulation model for the ow-induced vibrations in tube bundles.
The ow eld equations including turbulence were solved using
the boundary element method; however the tubes were structur-
ally treated as rigid cylinders supported by linear elastic strings.
The work-rate concept was also utilized by Yetisir et al. 16 to
investigate the possibility of establishing a damage criterion for
fretting wear based on the ow-induced vibration characteristics.
They developed a simple criterion based on parameters such as
vibration frequency, midspan vibration amplitude, span length,
tube mass, and an empirical wear coefcient. In Ref. 16, the
excitation was due to turbulence only, wherein uidelastic forces
were not modeled.
Fischer and Strohmeier 17 introduced a coupled uid-
structure interaction model to evaluate the stability of tube
bundles in crossow. A three-dimensional transient model is de-
veloped, which was augmented by a structural response model
based on beam theory and frictional impact. Reasonable results
were obtained for the case of xed-xed tube in crossow. The
model, however, does not address the effect of the tube pitch ratio
and tube arrangement in the bundle.
Au-Yang 18,19, in his sequel papers, reviewed the theoretical
development of the acceptance integral method to estimate the
random vibration of structures subjected to turbulent ow. The
closed-form and nite element solutions together with a standard
commercial nite element structural-analysis computer program
were employed to compute the joint and cross acceptances for
tubes and beams subjected to crossow. The results were pre-
sented in the form of charts, and steps were given to show how to
use these charts together with standard commercial nite element
structural-analysis programs to estimate the responses of single
and multispan tubes to crossow turbulence-induced vibration.
It has become evident that the modeling of the complex dynam-
ics of uidelastic forces that give rise to vibrations of tube bundles
requires a great deal of experimental insight. Experimental inves-
tigations on the phenomenon of ow-induced vibrations in heat
exchangers were recognized and pursued for the following two
reasons: a to gain more insight into the nature of such complex
dynamic behavior and b to reduce the complexity of the derived
mathematical models in light of some experimental ndings.
Grover and Weaver 20,21 presented a sequel of experimental
studies of crossow-induced vibration of the tube array, and
pointed out some observations over the range of their tests. Lever
and Weaver 22 made use of the ndings of such previous ex-
perimental investigations and developed a theoretical model for
the uidelastic instability in heat exchanger tube bundles. They
preceded their theoretical model derivation by a series of experi-
ments, thus paving the ground for considerable simplication of
the model. Based on experimental results, they concluded that
only a single elastic tube and the ow streams immediately adja-
cent to either side of the tube are required to model the essential
features of the uidelastic system. The developed model was in
very good agreement with the experimental results, and further
model improvement was recommended to include more sophisti-
cated uid mechanics, and to study the effect of tube pitch ratio
and pattern on the predicted stability.
To account for some uidelastic effects that could not be tack-
led by the quasisteady ow theory, Tanaka and co-workers
2325 introduced a method for calculating the critical ow ve-
locity based on the unsteady ow theory. They presented a dy-
namic model of a tube bundle in crossow, wherein the unsteady
uidelastic forces, for small tube vibrations, are represented by a
linear equation with three terms representing the added mass,
damping, and stiffness effects. The model addressed the coupling
between adjacent cylinders in the bundle; however each tube is
structurally modeled by a single degree-of-freedom lumped-
parameter system. Equations were presented in matrix form in-
cluding coupling, and experimental measurements were utilized to
determine the uid-induced force coefcients.
Granger et al. 26 used the same linear model of uidelastic
forces by Chen 8 to write the dynamic model for a bundle of N
tubes in crossow. The 2N degree-of-freedom model was later
approximated by a single degree-of-freedom system called the
global system. Experimental measurements were performed on a
tube bundle with several instrumented tubes, and the measured
modal parameters were then used to determine the global damping
and natural frequency, which in turn were used to determine the
uidelastic force coefcients. This approach, however, is depen-
dent on indirect identication techniques and the associated digital
signal processing method.
A methodology for modeling ow-induced vibrations of tubes
in crossow was presented by Chen and co-workers 27,28. The
unsteady ow theory is utilized in establishing the uidelastic
force coefcients. The uid-induced dynamic forces are then
added as exciting forces. Such forces were modeled by a similar
linear expression to that used by Tanaka et al. 24. The uid force
coefcients were stated analytically and justied by experiential
measurements. They concluded that uidelastic coefcients de-
pend on the tube arrangement, pitch, oscillation amplitude, re-
duced ow velocities, and Reynolds number.
Although considerable progress has been made in the area of
ow-induced vibrations since the early 1970, it remains necessary
to understand the ow-induced vibration mechanisms for all pos-
sible ow situations. To date, there are no accurate criteria by
which one could pinpoint the onset of uidelastic instability in
heat exchangers. The criteria set by TEMA 29 are relied on in
industry, though it is not adequate in predicting the onset of dam-
aging ow-induced vibrations in many situations. Pettigrew et al.
30 presented a comprehensive account of the problem of ow-
induced vibrations that still affects the performance and reliability
of heat exchangers. They concluded that there are still some im-
portant areas requiring further investigation, e.g., the prediction of
fretting wear and the understanding of damping mechanisms. Re-
cently Pettigrew and Taylor 31,32 presented an excellent over-
view of ow-induced vibrations in heat exchangers and outlined
some design guidelines to avoid the onset of such damaging vi-
brations.
Now, one must recognize that dynamic modeling in conjunction
with experimental measurements is the most feasible route to
tackle such a problem. Accordingly, a reliable dynamic model
must be tuned by the experimental evaluation of some uid-
induced force coefcients. However, most of the analytical mod-
els presented used simple lumped-parameter models to represent
the elastodynamic behavior of the tubes, wherein the uidelastic
coupling is congured for a single degree-of-freedom tube. Few
011301-2 / Vol. 131, FEBRUARY 2009 Transactions of the ASME
Downloaded 28 Nov 2008 to 212.26.1.29. Redistribution subject to ASME license or copyright; see http://www.asme.org/terms/Terms_Use.cfm
investigators utilized commercial nite element codes to model
tube bundles, with no mention of whether the uidelastic coupling
was integrated into such models.
In this paper, a uidelastic dynamic model using the nite ele-
ment approach is developed. In this context, each elastic tube is
modeled by a set of nite tube elements comprising a set of nodal
degrees-of-freedom. The resulting nonlinear dynamic model ad-
mits the unsteady uidelastic coupling in terms of experimentally
measured force coefcients. The developed nite element formu-
lation allows the uidelastic coupling to exist over the length of
the entire tube, thus avoiding the assumption adopted by several
previous investigations when the tube was treated as a single
lumped mass suspended by an elastic support. The resulting non-
linear eigenvalue problem is iteratively solved and the instability
conditions are determined. The experimental investigation needed
to measure the uidelastic force coefcients is presented in a com-
panion paper by the authors 33.
2 The Elastodynamic Model
Based on the actual heat exchanger construction, all the tubes
of the tube bundle were normally made of the same material and
had the same cross-sectional dimensions. The developed formula-
tion, however, is written in a general form to admit different geo-
metrical and material properties for each elastic component.
2.1 Assumptions and Basic Coordinates. The following are
the basic assumptions underlying the elastodynamic modeling: a
The material of the elastic tube is homogeneous and isotropic; b
the deection of the tube is produced by the displacement of
points of the centerline; and c the shear deformation for such
slender tube conguration is neglected. Now, let us consider an
element of the elastic tube, as shown in Fig. 1. The deformation of
a general point p in the cross section can be represented by the
position vector r
p
, which can be expressed in the form
r
p
= r
o
+ e 1
where r
o
is the unreformed position of point p, and e is the elastic
deformation vector, which can be written in the form e=Nq.
Now, Eq. 1 becomes
r
p
= r
o
+ Nq 2
where q is the vector of nodal coordinates and N is the shape
function matrix. The velocity vector is then obtained from Eq. 2
by differentiation with respect to time. Noting that the heat ex-
changer tube does not execute reference motion, the time deriva-
tive of Eq. 2 can be expressed as
r
p
= Nq 3
2.2 Kinetic Energy Expressions. In order to utilize the
Lagrangean approach, one needs to derive expressions for the
systems kinetic and potential energies. Using the velocity vector
in Eq. 3, the kinetic energy expression can be written as
KE =
1
2

s
r
p

T
r
p
dv 4
Using Eq. 3 into Eq. 4, one obtains
KE =
1
2

s
v
q
T
N
T
Nq dv =
1
2
q
T
Mq 5
where M =
v
N
T
Ndv is the elastic mass matrix that accounts
for axial, bending and torsional deformations of the tube element,
and
s
is the mass density of the tube material; details of the mass
matrix will be derived in the next section.
2.3 Strain Energy Expressions. There are six nodal coordi-
nates, which are dened as follows: uaxial deformation along
the x-direction, vtranslational deformation about the y-axis, w
translational deformation about the z-axis,
y
=vx, t / x
nodal bending rotation about the y-axis,
z
=wx, t / x
nodal bending rotation about the z-axis, and
x
nodal torsional
rotation about the x-axis Noting that the cross-sectional second
moment of area I
z
=I
y
=Ix, v/ x =
y
and w/ x =
z
, the
strain energy expression can be written in the form
SE =
1
2

0
l

EIx

y
x

2
+

z
x

2

+ EA
u
x

dx
+
1
2

0
l
GJ

2
dx 6
Using the standard nite element formulation, as explained in the
next section, Eq. 6 can be written in compact matrix form as
SE =
1
2
q
T
Kq 7
where K is a 1212 elemental stiffness matrix, which accounts
for bending, axial, and torsional stiffness, respectively.
Structural Dynamics FEM Formulation. In this formulation,
a nite two-node tube element, with six degrees-of-freedom per
node, is considered. Accordingly, the elemental nodal coordinate
vector, as shown in Fig. 2, can be written as
q = u
1
, v
1
, w
1
,
x
1
,
y
1
,
z
1
, u
2
, v
2
, w
2
,
x
2
,
y
2
,
z
2

T
8
Now, assuming a feasible displacement eld, the axial and bend-
ing deformations can be expressed in the form
Fig. 1 The generalized coordinate system
Fig. 2 Nodal coordinates of the tube element
Journal of Pressure Vessel Technology FEBRUARY 2009, Vol. 131 / 011301-3
Downloaded 28 Nov 2008 to 212.26.1.29. Redistribution subject to ASME license or copyright; see http://www.asme.org/terms/Terms_Use.cfm

ux, t
vx, t
wx, t

N
u1
0 0 0 0 0 N
u2
0 0 0 0 0
0 N
v1
0 0 N
v2
0 0 N
v3
0 0 N
v4
0
0 0 N
v1
N
v2
0 0 0 0 N
v3
N
v4
0 0

q
=

N
u
N
v
N
w

q = N
t
xq 9

=
0 N
1
0 0 N
2
0 0 N
3
0 0 N
4
0
0 0 N
1
N
2
0 0 0 0 N
3
N
4
0 0
q =
N
y
N
z
q = N

xq 10
x, t = 0 0 0 0 0 N
1
0 0 0 0 0 N
2
q = N

q 11
where N
u
, N
v
and N
w
are the translational shape functions, N
y
and N
z
are the elastic rotational shape functions, and N

is the
torsional deformation shape function. These shape functions were
derived by Perzemieniecki 34, or Bazoune et al. 35. The nite
element expressions of the coefcient matrices are given by
K =

0
l
B
e

T
EIB
e
dx +

0
l
B
a

T
EAB
a
dx +

0
l
B
a

T
EAB
a
dx
12
and
M =

0
l
N
v

s
AN
v
dx +

0
l
N

T
I
D
N

dx +

0
l
N

T
I
p
N

dx
13
where the three terms in Eq. 12 represent the bending, axial, and
torsional stiffness matrices, respectively. The derivatives of the
shape function matrices are given by B
e
=N

/ x, B
a

=N
u
/ x, and B

=N

/ x. In Eq. 13, the three terms rep-

resent the translational mass matrix, the rotational inertia mass
matrix, and the torsional mass matrix, respectively.
2.5 Equations of Motion. The dynamic equations of motion
that represent the elastodynamic behavior of the tube can be de-
rived using the Lagrangean approach. Denoting q
i
as the vector
of nodal displacements of tube i, one can substitute the
Lagrangean function in the variational form, carry out the associ-
ated differentiations, and then perform standard nite element as-
sembly procedure to express the equation of motion of tube i in
the following nal form:
M
i
q
i
+ D
i
q
i
+ K
i
q
i
= Q
i
14
Equation 14 represents the elastodynamic model of a tube ele-
ment, where the vector Q
i
is the generalized force vector that
may contain all externally applied forces, and may also contain
the time-dependent uidelastic coupling forces, if dynamic re-
sponse analysis is required. In this context, some subscripts are
introduced to identify the constituent matrices of a given tube as
M
s
i
q
i
+ D
s
i
q
i
+ K
s
i
q
i
= Q
i
15
For the rest of the derivation, the subscript s is introduced to refer
to the intrinsic structural properties, e.g., structural mass, damp-
ing, and stiffness properties, while the subscript f refers to the
corresponding uidelastic terms. The uid-structure interaction, as
manifested by the uidelastic forces, can be represented to include
the coupling between the adjacent tubes in the tube bundle. Let
the uidelastic force that couples tube i and tube j be
F
ij
= C
D
ij
s
ij
+ C
K
ij
s
ij
16
where s
ij
is a vector in the yz-plane, which represents the distance
between tube i and tube j, and C
D
ij
is the damping coefcient that
depends on the tube diameter, uid density, and the bundle gap
velocity. C
K
ij
is the stiffness coefcient that depends on the uid
density and the bundle gap velocity. Using the virtual work ex-
pression, one can write the work done by F
ij
as
W
F
ij
= F
ij
s
ij
= F
ij
s
ij
/q
i
q
i
F
ij
s
ij
/q
j
q
j
= Q
F
iT
q
i
+ Q
F
jT
q
j
17
The vectors Q
F
i
and Q
F
j
are the generalized forces associated with
the uidelastic coupling between tube i and tube j. These forces
are to be added to the right hand side of the equation of motion.
Now, Eq. 15 can be written for tubes i and j, including the
added-mass effect, as

M
i
0
0 M
j

q
i
q
j

D
s
i
0
0 D
s
j

q
i
q
j

K
s
i
0
0 K
s
j

q
i
q
j

Q
F
i
Q
F
j

18
where M is the mass matrix that includes both the structural iner-
tia properties and the added-mass effects, i.e., M
i
=M
s
i
+M
f
i
. The
added-mass matrix M
f
i
is a function uid density and the tube
dimensions. Equation 18 can be written in a general assembled
form to represent all active tubes in the tube array. The added-
mass effects, as well as the coupling uidelastic forces, are deter-
mined experimentally by estimating the uidelastic coefcients.
3 FEM Formulation of the Fluidelastic Forces
In this formulation, a crossow across a 33 square tube
bundle of three rows and three columns is considered. Previous
investigations have concluded that the uidelastic coupling for a
given tube is dominated by the adjacent tubes, and that the effect
of other distant rows may be insignicant 23. Nevertheless, the
presented formulation is directly applicable to a tube bundle of
any size. Let us consider a tube arrangement, as shown in Fig. 3.
In the case of multirow cylinders, the force on tube O is con-
sidered to be predominantly inuenced by the vibrations of the
upper U, the lower D, the right R, and the left L tubes. The
uid dynamic force acting on tube O is a function of the vibra-
tional motions of the tube itself and the adjacent tubes L, R, U,
and D, both in the x- and y-directions. The linearity assumption
of the augmented effect of the uid dynamic forces has been
veried by some experimental investigators, e.g., Tanaka et al.
011301-4 / Vol. 131, FEBRUARY 2009 Transactions of the ASME
Downloaded 28 Nov 2008 to 212.26.1.29. Redistribution subject to ASME license or copyright; see http://www.asme.org/terms/Terms_Use.cfm
25. Consequently, the superposition principle can be applied to
express the uidelastic effect as a linear combination of the u-
idelastic forces produced by different active tubes in the bundle as
follows:
F
y
=
1
2

f
V
2

k=1
5
C
yky
Y
k
+ C
ykz
Z
k
19
F
z
=
1
2

f
V
2

k=1
5
C
zky
Y
k
+ C
zkz
Z
k
20
Here, k=1, . . . , 5 corresponds to the tubes O, L, R, U, and H,
respectively. The subscripts y and z denote the vibration displace-
ments in the y- and z-directions, respectively. Following Tanakas
notations, each uid dynamic force coefcient is identied by
three subscripts. The rst sufx is associated with the direction of
the uid force, the second with the position of the vibrating tube,
and third with the direction of the tube vibration. In the case of the
square array, the following relations are obtained from the geo-
metrical properties and symmetry considerations of the array:
C
yRy
= C
yLy
, C
zRz
= C
zLz
, C
yRz
= C
yLz
, C
zRy
= C
zLy
C
yOz
= C
zOy
= C
yUz
= C
zUy
= C
yDz
= C
zDy
= 0
The following matrix explains the location of each tube with re-
spect to its surrounding tubes:
Tube Position U L O R D
Tube Number
1
2
3
4
5
6
7
8
9

0 0 1 2 4
0 1 2 3 5
0 2 3 0 6
1 0 4 5 7
2 4 5 6 8
3 5 6 0 9
4 0 7 8 0
5 7 8 9 0
6 8 9 0 0

21
Using the standard nite element assembly procedure, Eq. 14
can be written for all tubes considered in the active bundle. The
stiffness, mass, and damping matrices due to uid coupling are
added to the respective structural stiffness, mass, and damping
matrices of the bundle to obtain the augmented equation of motion
as
Mq + Dq + Kq = Q 22
where M =M
s
+M
f
, K =K
s
+K
f
, and D =D
s

+D
f
are the assembled nite element coefcient matrices of the
whole bundle. In this context, the matrix K
s
is the structural
stiffness matrix of the bundle and K
f
is the corresponding added
stiffness matrix due to uidelastic forces. The matrix D
s
is struc-
tural damping matrix of the whole bundle, which represents the
material damping and is commonly represented by a proportional
damping model in terms of a linear combination of both the struc-
tural stiffens and mass matrices. In this formulation, the spanwise
effect on the uidelastic forces is taken into consideration, and
consequently the model is capable of addressing the three-
dimensional instabilities 36. The nite element expressions of
the uid-induced mass, stiffness, and damping matrices M
f
,
K
f
, and D
f
are assembled from the constituent 99 coupling
matrices, which are established at each nodal plane of the bundle
as follows:
M

f
=

M
1/1
M
1/2
0 M
1/4
0 0 0 0 0
M
2/1
M
2/2
M
2/3
0 M
2/5
0 0 0 0
0 M
3/2
M
3/3
0 0 M
3/6
0 0 0
M
4/1
0 0 M
4/4
M
4/5
0 M
4/7
0 0
0 M
5/2
0 M
5/4
M
5/5
M
5/6
0 M
5/8
0
0 0 M
6/3
0 M
6/5
M
6/6
0 0 M
6/9

0 0 0 M
7/4
0 0 M
7/7
M
7/8
0
0 0 0 0 M
8/5
0 M
8/7
M
8/8
M
8/9

0 0 0 0 0 M
9/6
0 M
9/8
M
9/9

23
Fig. 3 Instrumented tube array arrangement
Journal of Pressure Vessel Technology FEBRUARY 2009, Vol. 131 / 011301-5
Downloaded 28 Nov 2008 to 212.26.1.29. Redistribution subject to ASME license or copyright; see http://www.asme.org/terms/Terms_Use.cfm
K

f
=

K
1/1
K
1/2
0 K
1/4
0 0 0 0 0
K
2/1
K
2/2
K
2/3
0 K
2/5
0 0 0 0
0 K
3/2
K
3/3
0 0 K
3/6
0 0 0
K
4/1
0 0 K
4/4
K
4/5
0 K
4/7
0 0
0 K
5/2
0 K
5/4
K
5/5
K
5/6
0 K
5/8
0
0 0 K
6/3
0 K
6/5
K
6/6
0 0 K
6/9

0 0 0 K
7/4
0 0 K
7/7
K
7/8
0
0 0 0 0 K
8/5
0 K
8/7
K
8/8
K
8/9

0 0 0 0 0 K
9/6
0 K
9/8
K
9/9

24
D

f
=

D
1/1
D
1/2
0 D
1/4
0 0 0 0 0
D
2/1
D
2/2
D
2/3
0 D
2/5
0 0 0 0
0 D
3/2
D
3/3
0 0 D
3/6
0 0 0
D
4/1
0 0 D
4/4
D
4/5
0 D
4/7
0 0
0 D
5/2
0 D
5/4
D
5/5
D
5/6
0 D
5/8
0
0 0 D
6/3
0 D
6/5
D
6/6
0 0 D
6/9

0 0 0 D
7/4
0 0 D
7/7
D
7/8
0
0 0 0 0 D
8/5
0 D
8/7
D
8/8
D
8/9

0 0 0 0 0 D
9/6
0 D
9/8
D
9/9

25
where
K
i/j
=
1
2

f
d
2

2
C
O
cos
i/j
V
r
2
2
3
C
mi/j
/4
2
26
M
i/j
=

4

f
d
2
C
mi/j
27
D
i/j
=
1
8
2

f
d
2
V
r
2
C
O
sin
i/j
28
C
O
cos
i/j
= Block DiagonalA
i/j
0 A
i/j
0 A
i/j
0
29
where 0 is a 22 null matrix, and A
i/j
is given by
A
i/j
=
C
O
cos
YKY
C
O
cos
YKZ
C
O
cos
ZKY
C
O
cos
ZKZ
30
The subscript k refers to O, L, R, U, or H, depending on the
position of tube j with respect to tube i. The constituent coupling
matrices in Eq. 2325 are assembled into the associated coef-
cient matrices according to the placement array of Eq. 21. For
example, if i =1 and j =2, then k stands for R because tube 2 is
positioned to the right of tube 1, i.e.,
A
1/2
=
C
O
cos
YRY
C
O
cos
YRZ
C
O
cos
ZRY
C
O
cos
ZRZ
31
Here, V is the crossow velocity,
f
is the uid density, d is the
outer diameter of the tube, and V
R
is the reduced velocity. The
coefcient submatrices M
f
, K
f
, and D
f
are the elemental
added-mass, stiffness, and damping matrices on tube i due to u-
idelastic coupling with tube j. It is noted that when i = j, such
matrices account for the uid-induced effects due to the vibration
of the tube itself.
4 The Eigenvalue and Stability Analysis
Assuming no externally applied forces, the uidelastic forces
on the right side of Eq. 22 can be absorbed into the left side, thus
leading to the representation of the nal dynamic model in the
following compact form:
M

q + D

q + K

q = 0 32
where M

, D

, and K

are the global mass, damping, and stiff-

ness matrices that include both structural and uidelastic coef-
cients for the whole tube bundle. The eigenvalue problem associ-
ated with this nonself adjoint form can be established by writing
Eq. 32 in the state space form as
d
dt

q
q

=

0 I
M

1
K

1
D

q
q

33
or simply as
y = Ay 34
where y =q
T
q
T

T
is the state vector, and A =A is the
state coefcient matrix, which is a function of the natural fre-
quency of the tube bundle. A is a general matrix, thus the result-
ing eigenvalues are complex. The dependence of A on the natu-
ral frequency results in a nonlinear eigenvalue problem. Neumaier
37 presented an inverse iteration scheme for the nonlinear eigen-
value problem. A class of nonlinear eigenvalue problems encoun-
tered in solid-structure problems has been addressed by Conca et
al. 38.
The scheme adopted in this paper employs an inverse iteration
outer loop with the MATLAB complex eigenvalue solver as its inner
core. The method rst calculates structural stiffness, mass, and
damping matrices for the tube bundle. It also interpolates the uid
force coefcients by curve tting so that these coefcients can be
determined at any iteration step of the reduced velocity. It then
updates the stiffness, mass, and damping matrices with the current
value of uidelastic effects. The modal characteristics of the
whole system tube bundle are then determined by solving the
associated complex eigenvalue problem. The critical reduced ve-
011301-6 / Vol. 131, FEBRUARY 2009 Transactions of the ASME
Downloaded 28 Nov 2008 to 212.26.1.29. Redistribution subject to ASME license or copyright; see http://www.asme.org/terms/Terms_Use.cfm
locity is dened as the reduced velocity at the onset of instability.
The iteration is indexed over the reduced velocity, and the itera-
tion is terminated once the critical reduced velocity is reached. In
other words, the critical reduced velocity is one at which the high-
est real part of an eigenvalue changes its sign from negative
stable region to positive unstable region.
5 Numerical Results and Conclusions
In order to test the developed numerical scheme, the data of the
test example reported by Tanaka and co-workers 23,25 are used.
The case consists of a 34 square array with a pitch-to-diameter
ratio of 1.33. Tubes 30 mm 1.18 in. in diameter and 300 mm
11.8 in. in length were used. For this case study, Tanaka et al.
25 tabulated the uidelastic coefcients, which they have ex-
perimentally determined using their experimental setup.
The developed scheme was invoked by utilizing the aforemen-
tioned available test data. In this regard, the nonlinear eigenvalue
problem was generated and solved numerically to study the effect
of different tube arrays and ow parameters on the onset of ow-
induced vibrations. Using the same mass ratio and reduced veloc-
ity denitions of Tanaka, Table 1 displays the results obtained
using the developed nite element model FEM formulation for
the case of having the same mass damping parameter of 0.19.
Table 2 shows the comparison of results for a xed mass ratio of
1.852 and a range of damping values. As depicted in Tables 1 and
2, the results obtained by the developed nite element model are
very comparable to those obtained by Tanaka; however they were
consistently slightly lower. This may be anticipated, which could
be attributed to two reasons, namely, a the frequency calcula-
tions and b the uidelastic forces calculations. Firstly, the single
lumped-mass model assumed by Tanaka is known to underesti-
mate the calculated natural frequencies, while the consistent mass
FEM formulation overestimates calculated frequencies. Secondly,
the uidelastic coupling in Tanakas model is associated with a
single lumped-mass, while in the nite element it is a function of
a set of nodal degrees-of-freedom associated with the distributed
mass over the whole tube length.
The developed numerical scheme is then used to study the ef-
fect of the pitch-diameter-ratio on the onset of uidelastic
instability, as marked by the recoded value of the critical velocity.
Figure 4 shows the results for the case when =2 and =0.2,
while varied in the range 1.22.5. The numerical prediction of
the critical velocity showed that the value of critical velocity in-
creased as increased. However, for values of above 1.75, the
increase in critical velocity is relatively negligible. A similar trend
was reported by Tanaka et al. 24, in which they pointed out that
the critical velocity did not vary for greater than 2.
As manifested by the aforementioned comparisons, the devel-
oped nite element numerical scheme demonstrates a good degree
of accuracy in predicting the onset of instability associated with
the ow-induced vibrations in tubular heat exchangers. The devel-
oped scheme has the advantage of representing the uidelastic
coupling forces in terms of a set of degrees-of-freedom distributed
over the entire tube length, thus providing a more accurate pre-
diction of such forces.
Acknowledgment
This research work is funded by Saudi Aramco, Project No.
ME2212. The authors greatly appreciate the support provided by
Saudi Aramco and King Fahd University of Petroleum & Minerals
during this research.
Table 1 Critical velocity estimates =0.19
a
Mass ratio
b

Critical velocity
Ref. 24 Present method
2.22 2.024 1.983
4.44 2.283 2.204
8.88 2.394 2.311
17.76 2.448 2.398
35.52 2.474 2.421
71.04 2.485 2.442
a
=m/
f
d is the mass damping parameter, where is the logarithmic decrement, m
is the mass per unit length of the tube,
f
is the uid density, and d is the tube outer
diameter.
b
Mass ratio: =m/
f
d
2
.
Table 2 Critical velocity estimates =1.852
Logarithmic
decrement

Critical velocity
Ref. 25 Present method
0.0023 1.35 1.291
0.0061 1.45 1.387
0.0297 1.75 1.696
0.033 1.80 1.734
0.087 2.35 2.221
0.166 2.70 2.586
Fig. 4 Effect of pitch-diameter-ratio
Journal of Pressure Vessel Technology FEBRUARY 2009, Vol. 131 / 011301-7
Downloaded 28 Nov 2008 to 212.26.1.29. Redistribution subject to ASME license or copyright; see http://www.asme.org/terms/Terms_Use.cfm
Nomenclature
A cross-sectional area of tube
C
ijk
uid dynamic force coefcient
C
m
added-mass coefcient
C
D
damping coefcient
C
K
stiffness coefcient
C
0
constant amplitude
F force
G shear modulus
d tube diameter
D damping matrix
K stiffness matrix
M mass matrix
N shape function matrix
q nodal coordinate vector
Q generalized force vector
r
p
position vector
V
r
reduced velocity

s
density of tube material

f
density of uid
torsional deformation
phase difference
frequency rad/s
References
1 Wallis, R. P., 1939, Photographic Study of Fluid Flow Between Banks of
Tubes, Engineering London, 148, pp. 423426.
2 Putnam, A. A., 1959, Flow Induced Noise in Heat Exchangers, ASME J.
Eng. Power, 814, pp. 417422.
3 Paidoussis, M. P., 1983, A Review of Flow-Induced Vibrations in Reactor and
Reactor Components, Nucl. Eng. Des., 74, pp. 3160.
4 Price, S. F., 1995, A Review of Theoretical Models for Fluid-Elastic Instabil-
ity of Cylinder Arrays in Crossow, J. Fluids Struct., 9, pp. 463518.
5 Weaver, D. S., Ziada, S., Au-Yang, M. K., Chen, S. S., Paidoussis, M. P., and
Pettigrew, M. J., 2000, Flow-Induced Vibrations in Power and Process Plant
Components-Progress and Prospects, ASME J. Pressure Vessel Technol.,
122, pp. 339348.
6 Chen, S., 1987, Flow-Induced Vibrations of Circular Cylindrical Structures,
Hemisphere, New York.
7 Katinas, V., and Zukauskas, A., 1997, Vibrations of Tubes in Heat Exchangers,
Begell House, NY.
8 Chen, S. S., 1983, Instability Mechanisms and Stability Criteria of a Group of
Circular Cylinders Subjected to Cross-FlowPart 1: Theory, ASME J. Vib.,
Acoust., Stress, Reliab. Des., 1051, pp. 3139.
9 Chen, S. S., 1983, Instability Mechanisms and Stability Criteria of a Group of
Circular Cylinders Subjected to Cross-FlowPart 2: Numerical Results and
Discussions, ASME J. Vib., Acoust., Stress, Reliab. Des., 105, pp. 253260.
10 Cai, Y., Chen, S. S., and Chandra, S., 1992, A Theory for Fluidelastic Insta-
bility of Tube-Support Plate Inactive Modes, ASME J. Pressure Vessel Tech-
nol., 114, pp. 139148.
11 Cai, Y., and Chen, S. S., 1993, Chaotic Vibrations of Nonlinearly Supported
Tubes in Crossow, ASME J. Pressure Vessel Technol., 115, pp. 128134.
12 Eisinger, F. L., Rao, M. S. M., Steininger, D. A., and Haslinger, K. H., 1995,
Numerical Simulation of Cross-Flow-Induced Fluidelastic Results, ASME J.
Pressure Vessel Technol., 117, pp. 3139.
13 Granger, S., and Paidoussis, M. P., 1996, An Improvement to the Quasi-
Steady Model With Application to Cross-Flow-Induced Vibration of Tube Ar-
rays, J. Fluid Mech., 320, pp. 163184.
14 Ichioka, T., Kawata, Y., Nakamura, T., Izumi, H., Kobayashi, T., and Taka-
matsu, H., 1997, Research on Fluid Elastic Vibration of Cylinder Arrays by
Computational Fluid Dynamics: Analysis of Two Cylinders and a Cylinder
Row, JSME Int. J., Ser. B, 401, pp. 1624.
15 Kassera, V., and Strohmeier, K., 1997, Simulation of Tube Bundle Vibrations
Induced by Cross-Flow, J. Fluids Struct., 11, pp. 909928.
16 Yetisir, M., Mckerrow, E., and Pettigrew, M. J., 1998, Fretting Wear Damage
of Heat Exchanger Tubes: A Proposed Damage Criterion Based on Tube Vi-
bration Response, ASME J. Pressure Vessel Technol., 120, pp. 297305.
17 Fischer, M., and Strohmeier, K., 1999, Three-Dimensional Simulation of
Single Fixed-Fixed Tube Vibration Induced by Cross-Flow Including Bending
and Torsion, Emerging Technologies in Fluids, Structures, and Fluid/
Structure Interactions, Boston, MA, ASME Vol. PVP-396, pp. 285290.
18 Au-Yang, M. K., 2000, Joint and Cross Acceptances for Cross-Flow-Induced
VibrationPart I: Theoretical and Finite Element Formulations, ASME J.
Pressure Vessel Technol., 122, pp. 349354.
19 Au-Yang, M. K., 2000, Joint and Cross Acceptances for Cross-Flow-Induced
VibrationPart II: Charts and Applications, ASME J. Pressure Vessel Tech-
nol., 122, pp. 355360.
20 Grover, L. K., and Weaver, D. S., 1978, Cross-Flow Induced Vibration in a
Tube Bank-Vortex Shedding, J. Sound Vib., 59, pp. 263276.
21 Weaver, D. S., and Grover, L. K., 1978, Cross-Flow Induced Vibration in a
Tube Bank-Turbulent Buffeting and Fluid Elastic Instability, J. Sound Vib.,
59, pp. 277294.
22 Lever, J. H., and Weaver, D. S., 1982, A Theoretical Model for Fluid-Elastic
Instability in Heat Exchanger Tube Bundles, ASME J. Pressure Vessel Tech-
nol., 104, pp. 104147.
23 Tanaka, H., and Takahara, S., 1981, Fluid Elastic Vibrations of Tube Array in
Cross Flow, Sound Vib., 771, pp. 1937.
24 Tanaka, H., Takahara, S., and Ohta, K., 1982, Flow-Induced Vibrations in
Tube Arrays with Various Pitch-to-Diameter Ratios, ASME J. Pressure Vessel
Technol., 104, pp. 168174.
25 Tanaka, H., Tanaka, K., Shimiuz, F., and Takahara, S., 2002, Fluidelastic
Analysis of Tube Bundle Vibrations in Cross-ow, J. Fluids Struct., 161,
pp. 93112.
26 Granger, S., Campistron, R., and Lebret, J., 1993, Motion-Dependent Excita-
tion Mechanisms in a Square In-Line Tube Bundle Subject to Wear Cross-
Flow: An Experimental Modal Analysis, J. Fluids Struct., 7, pp. 521550.
27 Chen, S. S., Zhu, S., and Jendrzejczyk, J. A., 1994, Fluid Damping and Fluid
Stiffness of a Tube in Crossow, ASME J. Pressure Vessel Technol., 116, pp.
370383.
28 Chen, S. S., Cai, Y., and Zhu, S., 1996, Flow-Induced Vibration of Tubes in
Cross-Flow, ASME J. Offshore Mech. Arct. Eng., 118, pp. 253258.
29 TEMA Standards of Heat Exchanger Manufacturer Association, Mechanical
Standards TEMA Class RCB, Section, 5, pp. 34.
30 Pettigrew, M. J., Taylor, C. E., Fisher, N. J., Yetisir, M., and Smith, B. W. A.,
1998, Flow-Induced Vibration: Recent Findings and Open Questions, Nucl.
Eng. Des., 185, pp. 249276.
31 Pettigrew, M. J., and Taylor, C. E.N, 2003, Vibration Analysis of Shell-And-
Tube Heat Exchangers: An Overview, Part I: Flow, Damping, Fluidelastic
Instability, J. Fluids Struct., 18, pp. 469483.
32 Pettigrew, M. J., and Taylor, C. E., 2003, Vibration Analysis of Shell-And-
Tube Heat Exchangers: An Overview, Part II: Vibration Response, Fretting-
Wear, Guidelines, J. Fluids Struct., 18, pp. 485500.
33 Al-Kaabi, S. A., Khulief, Y. A., and Said, S. A., 2009, Prediction of Flow-
Induced Vibrations in Tubular Heat ExchangersPart II: Experimental Inves-
tigation, ASME J. Pressure Vessel Technol., 131, p. 011302.
34 Przemieniecki, J. S., 1968, Matrix Structural Analysis, McGraw-Hill, New
York.
35 Bazoune, A., Stephen, N. G., and Khulief, Y. A., 2003, Shape Functions for
ThreeDimensional Timoshenko Beam Elements, J. Sound Vib., 2592, pp.
473480.
36 Kevlahan, N. K., and Wadsley, J., 2005, Suppression of Three-Dimensional
Flow Instabilities in Tube Bundles, J. Fluids Struct., 20, pp. 611620.
37 Neumaier, A., 1985, Residual Inverse Iteration for the Nonlinear Eigenvalue
Problem,SIAM Soc. Ind. Appl. Math. J. Numer. Anal., 22, pp. 914923.
38 Conca, C., Planchard, J., and Vanninathan, M., 1989, Existence and Location
of Eigenvalues for Fluid-Solid Structures, Comput. Methods Appl. Mech.
Eng., 77, pp. 253291.
011301-8 / Vol. 131, FEBRUARY 2009 Transactions of the ASME
Downloaded 28 Nov 2008 to 212.26.1.29. Redistribution subject to ASME license or copyright; see http://www.asme.org/terms/Terms_Use.cfm