Damping of Heat Exchanger Tubes in rameters.

Semi-empirical formulations for damping of heat ex-

changer tubes in liquids are recommended for design purposes.
Liquids: Review and Design
Guidelines 2 Energy Dissipation Mechanisms
Generally, there are several possible energy dissipation mecha-
M. J. Pettigrew nisms that could contribute to tube damping, namely:
Fellow ASME • Tube related
BWC/AECL/NSERC Chair of Fluid-Structure Interaction 共a兲 internal or material damping;
Department of Mechanical Engineering, 共b兲 viscous damping between the tube and fluid;
Ecole Polytechnique, 共c兲 flow dependent damping due to the fluid flow around the
Montreal, QC H3T 1J4, Canada tube;
共d兲 damping due to the fluid flow inside the tube;
R. J. Rogers 共e兲 damping due to two-phase flow
Department of Mechanical Engineering,
• Support related
University of New Brunswick,
Fredericton, NB E3B 5A3, Canada 共a兲 friction at the tube-to-tubesheet joints;
共b兲 sliding friction between the tubes and tube-supports;
F. Axisa 共c兲 energy dissipated by tube-to-support impact and resulting
Vibration Consultant traveling waves;
11 Villa Poirier, 共d兲 squeeze-film damping in the clearance between the tubes
and supports;
Paris 75015, France 共e兲 viscous shear damping between the tubes and supports

As observed by Goyder 关1兴 there are two principal types of tube

This paper addresses the question of damping of multispan heat motion at the support: rocking and lateral motions, as shown in
exchanger tubes with liquids (mostly water) on the shell side. The Fig. 1. The motion at any given support may be a combination of
different energy dissipation mechanisms that contribute to damp- rocking and lateral motions.
ing are investigated. The available experimental data from the Rocking-type motion is predominant in the lower vibration
literature and from our own measurements are reviewed and ana- modes. Damping due to rocking motion is likely to be less than
lyzed. Three important energy dissipation mechanisms emerge. for lateral motion and thus may be more relevant in practice.
These are viscous damping between the tube and liquid, squeeze- The dynamic interaction between tubes and tube-supports may
film damping in the clearance between the tube, and support and be characterized in three main types, namely, sliding, impacting,
friction damping at the support. Viscous damping only accounts and scuffing, which is impacting at an angle followed by sliding
for approximately 25% of the total damping of a typical tube. 共Fig. 2兲. In practice, the dynamic interaction between a tube and
Thus, about 75% of the damping energy is dissipated at the sup- each tube-support is likely to be a combination of the above. The
port. Squeeze-film damping appears to be the most important en- energy dissipated by sliding is due to friction and is related to the
ergy dissipation mechanism. Squeeze-film damping is related to product of contact force and displacement. The energy dissipated
the support width and is inversely proportional to the tube fre- by impact is in the form of local deformation of the support,
quency. Damping is formulated in terms of tube and tube-support followed by stress wave propagation in the support and local de-
parameters. Semi-empirical formulations for damping of heat ex- formation of the tube, followed by high frequency traveling waves
changer tubes in liquids are recommended for design purposes. in the tube.
关DOI: 10.1115/1.4000711兴 Material damping is very small in heat exchanger tubing.
Haslinger and Thompson 关2兴 measured damping ratios of roughly
0.01% on tubes welded or brazed at both ends with no intermedi-
ate support. For such welded tubes, material damping is dominant
1 Introduction since the contribution from other energy dissipation mechanisms
Several flow-induced vibration excitation mechanisms may lead is minimal. Thus, the contribution of material damping may be
to severe vibration in heat exchangers and other process equip- neglected.
ments. These mechanisms are: fluidelastic instability, periodic Since viscosity and density are very small in air and light gases,
wake shedding, turbulence excitation, and acoustic resonance. Ex- tube-to-fluid viscous damping, squeeze-film damping, and viscous
cessive flow-induced vibration may cause heat exchanger tube shear damping are not significant in heat exchangers with gas on
failures either by fatigue or fretting-wear. To avoid such problems, the shell side. These damping mechanisms may be appreciable in
a thorough flow-induced vibration analysis considering all of the heavy gases 共e.g., high pressure gas兲 and are very important in
above vibration excitation mechanisms should be carried out at liquids.
the design stage. Such analyses require information about tube Liquid flow through a tube bundle will affect damping because
damping. of flow-dependent damping and fluidelastic forces. However, it is
This paper addresses the question of damping of multispan heat very difficult to separate energy dissipating fluid forces from flu-
exchanger tubes with liquids 共mostly water兲 on the shell side. The idelastic excitation forces in experiments with flexible tube
different energy dissipation mechanisms that contribute to damp- bundles. There is limited information on flow-dependent damping
ing are discussed. The available experimental data from the litera- in tube arrays 关3兴. They show that damping is not too dependent of
ture and from our own measurements are reviewed and analyzed. flow velocity below the critical velocity for fluidelastic instability.
Damping is formulated in terms of the tube and tube-support pa- In this paper the flow-dependent damping forces are considered
together with the fluidelastic instability forces.
Some researchers describe the decrease in damping at flow ve-
Contributed by the Pressure Vessel and Piping Division of ASME for publication
in the JOURNAL OF PRESSURE VESSEL TECHNOLOGY. Manuscript received September 11,
locities close to critical flow velocities for fluidelastic instability
2008; final manuscript received October 6, 2009; published online January 21, 2011. as negative damping. Instead, we consider that the apparent reduc-
Editor: G. E. Otto Widera. tion in damping is part of the fluidelastic instability phenomenon.

Journal of Pressure Vessel Technology Copyright © 2010 by ASME FEBRUARY 2011, Vol. 133 / 014002-1

Downloaded From: https://pressurevesseltech.asmedigitalcollection.asme.org on 06/18/2019 Terms of Use: http://www.asme.org/about-asme/terms-of-use

Damping due to a two-phase mixture surrounding the tubes can
be very significant 关6–9兴. This situation exists in many heat ex-
changers such as boilers, condensers, and nuclear steam genera-
tors. Two-phase damping depends on parameters such as mass
flux, void fraction, flow path geometry, and flow regime. This
topic is discussed in detail in Ref. 关9兴.

Fig. 1 Types of tube motion at the support location 3 Approach

Heat exchanger tube dynamics is inherently a nonlinear phe-
nomenon since it depends largely on the dynamic interaction be-
tween tubes and tube-supports. This is particularly so for squeeze-
Fluid flow inside a tube is unlikely to affect damping unless the film and friction-type dampings that depend entirely on the
flow velocity is very high 共i.e., approximately that required for relative motion between tubes and tube-supports. The analysis of
axial flow fluidelastic instability兲. Such internal flow velocities are this phenomenon would require a time domain nonlinear simula-
not normally encountered in heat exchanger tubes. However, two- tion of the tube dynamics, in which the details of sliding and
phase flow or sloshing due to partially filled tubes would contrib- impacting between tubes and tube-supports are modeled. Unfor-
ute to damping. Scriven and Hopley 关4兴 and Gravelle et al. 关5兴 tunately, we are still lacking some of the detailed information
studied this phenomenon. This source of damping depends very required to model the dynamic interaction between tubes and
much on flow regime and void fraction. It would be risky to rely tube-supports. Furthermore, the required nonlinear analysis is dif-
on the contribution of this energy dissipation mechanism unless ficult and is not yet generally available to the designer. Some
we have specific data for the situation being analyzed. progresses have been made in this area with the development of
Friction due to the relative motion between a tube and its tube- codes such as VIBIC 关10兴, H3DMAP 关11兴, SGFW-PC 关12兴, to predict
supports is a significant energy dissipation mechanism. In liquids, fretting-wear of heat exchanger tubes. These codes could be
there may be some lubrication between tubes and tube-supports, adapted to predict damping of heat exchanger tubes.
preventing metal-to-metal contact. This could alter the nature of However, for the time being, most designers are limited to
the friction forces. On the other hand, there should be some vis- quasi-linear vibration analyses, for which they have to obtain an
cous shear damping forces due to sliding between tubes and tube- equivalent linear damping value corresponding to realistic vibra-
supports. The exact nature of sliding forces in liquids is not yet tion amplitudes 关13,14兴. At first sight this approach may appear
well understood. contradictory since the diametral clearance between tubes and
At present, we do not know how to separate the energy dissi- supports is in the range of 0.25–0.8 mm. However, in well-
pated by tube impact and the resulting traveling waves from the designed heat exchangers, the tube vibration response at midspan
other damping mechanisms. Impact damping may be much less in is mostly less than 100 ␮m rms and much less at the supports
liquids than in gases. Impact forces are probably attenuated and 共usually less than 25 ␮m rms兲. This is significantly less than the
their contact rise time lengthened by squeeze-film forces. This available diametral clearance. Thus, the tubes do not generally
would tend to reduce higher frequency traveling waves. Thus, we vibrate back and forth across the available diametral clearance.
do not consider impact damping as a separate energy dissipation Instead, most tubes are not centered within the supports and are,
mechanism. It may be considered together with friction damping therefore, contacting or vibrating very close to one side of the
or squeeze-film damping. supports. Given some tube out-of-straightness and tube-support
In summary, tube-to-fluid viscous damping, squeeze-film damp- misalignment, it is difficult to imagine that many tubes of typi-
ing, and friction damping are the most important energy dissipa- cally 1 m span length would be located concentrically within, say,
tion mechanisms in liquids. Tube-to-fluid viscous damping is dis- a 0.40 mm diametral clearance without touching the support. The
cussed in Sec. 4. As will be seen later, squeeze-film damping in chances of a tube not touching a support are very small. In sum-
the clearance between tube and support is particularly important. mary, the vibration response is significantly larger than the motion
The formulation of this mechanism is developed in Sec. 6. Fric- at the supports and the system may be modeled as a quasi-linear
tion and viscous shear damping is discussed in detail in Sec. 7. system with equivalent linear damping over the small range of
amplitude of interest. Thus, it is reasonable to assume pinned
conditions to allow for a quasi-linear analysis to calculate the
vibration response.
The objective of this paper is to recommend appropriate damp-
ing values based on available data. There are a few publications
devoted to heat exchanger tube damping 关6,15–29兴. In addition,
there are some useful damping information buried in a number of
other publications related to heat exchanger vibrations, as listed in
the references 关30–39兴. Most of this information are qualified as
being tentative or preliminary. This is not surprising, considering
the nonlinear nature of the problem and the difficulties inherent in
measuring damping.
Our approach is to review the available data and formulate a
conservative but realistic minimum damping criterion based on
the three main energy dissipation mechanisms discussed earlier.
We propose to take the lower decile of all the available data as a
minimum damping level. The lower decile means that 10% of the
data is below the minimum damping criterion. This is reasonable
since the smaller damping values measured are usually the more
reliable and since vibration problems are usually associated with
Fig. 2 Types of dynamic interaction between the tube and the lower damping values. The higher damping values reported
tube-support may be due to significant preloads at the supports, caused by

014002-2 / Vol. 133, FEBRUARY 2011 Transactions of the ASME

Downloaded From: https://pressurevesseltech.asmedigitalcollection.asme.org on 06/18/2019 Terms of Use: http://www.asme.org/about-asme/terms-of-use

The damping data extracted from the available references is
outlined in Fig. 3 and compiled in the form of a database. A
sample of the database is shown in Table 1. This data is for mul-
tispan heat exchanger tubes in water. There are field data on real
heat exchanger tube bundles and laboratory data on single tubes in
the references listed. The tube configurations and experimental
considerations are outlined in Table 2. This damping data ex-
pressed in terms of the damping ratio ␨ is generally higher than
0.6% and appears dependent on frequency. The damping ratio is
the ratio of the damping coefficient C over the critical damping
coefficient Cc. The data were analyzed to find trends and to for-
mulate a damping criterion in terms of tube and tube-support pa-
rameters.

4 Tube-to-Fluid Viscous Damping

Of the three main energy dissipation mechanisms, viscous
damping between tube and fluid is the easiest to formulate. In still
fluid, the contribution of viscous fluid forces to damping is related
to the Stokes Number S or ␲ fD2 / 2␯, and to the degree of con-
finement of the heat exchanger tube within the tube bundle. Vis-
cous damping is a function of the Stokes Number, as shown in Eq.
共1兲
Fig. 3 Damping data for multispan heat exchanger tubes in
water
␨ v⬁ 冉 冊
␲ fD2
2␯
−0.5
共1兲

where ␨V is the viscous damping ratio, ␯ is the kinematic viscosity,

slightly bent tubes or misaligned supports. Anyway, the higher f is the vibration frequency, and D is the tube diameter. This
damping values are of no concern here since they do not lead to implies that viscous damping is inversely proportional to the
vibration problems. square root of frequency. Chen et al. 关28兴 developed a comprehen-

Table 1 Sample database page on heat exchanger tube damping in liquids „water…

Span
length Damping

Tube Total Clear. Total Viscous Support

Freq. f Diam. D Wall t mass mt mass m ᐉ1 ᐉ2 ᐉ3 Ds − D Thick. L ␨T ␨V ␨S Spans Mode
Data No. 共Hz兲 共mm兲 共mm兲 共kg/m兲 共kg/m兲 共mm兲 共mm兲 共mm兲 共mm兲 共mm兲 P / Da 共%兲 共%兲 共%兲 N No.

1 30.5 15.9 3.2 0.92 1.2 1170 1070 558 0.25 12.7 ⬁ 2 0.21 1.79 6 1
2 30.5 15.9 3.2 0.92 1.2 1170 1070 558 0.25 19 ⬁ 11 0.21 10.79 6 1
3 30.5 15.9 3.2 0.92 1.2 1170 1070 558 0.25 38.1 ⬁ 2.8 0.21 2.59 6 1
4 40.3 15.9 3.2 0.92 1.2 1170 1070 558 0.76 38.1 ⬁ 8 0.18 7.82 6
5 31.2 15.9 3.2 0.92 1.2 1170 1070 558 0.25 38.1 ⬁ 5 0.21 4.79 6 1
6 30.6 15.9 3.2 0.92 1.2 1170 1070 558 0.51 38.1 ⬁ 2 0.21 1.79 6 1
7 33.1 15.9 3.2 0.92 1.2 1170 1070 558 0.76 38.1 ⬁ 4 0.20 3.80 6
8 41.5 15.9 3.2 0.92 1.2 1170 1070 558 0.25 38.1 ⬁ 8 0.18 7.82 6
9 42.5 15.9 3.2 0.92 1.2 1170 1070 558 0.51 38.1 ⬁ 4 0.18 3.82 6
10 64.2 15.9 3.2 0.92 1.2 734 734 734 0.25 38.1 ⬁ 2 0.15 1.85 7
11 30 15.9 3.2 0.92 1.2 ⬁ 2.6 0.21 2.39
12 37 15.9 3.2 0.92 1.2 ⬁ 2.4 0.19 2.21
13 45 15.9 3.2 0.92 1.2 ⬁ 1.7 0.18 1.52
14 44.4 19 1.57 0.755 1.202 1450 1450 1450 0.8 12.7 1.25 1.6 0.43 1.34 8
15 33.7 19 1.57 0.755 1.202 1450 1450 1450 0.8 12.7 1.25 2.5 0.49 2.30 8
16 39.8 13 1.04 0.39 0.566 890 890 0 0.4 25.4 1.57 2.16 0.45 1.71 2
17 63.5 13 1.04 0.39 0.566 890 890 0 0.4 25.4 1.57 1.15 0.36 0.79 2
18 119 13 1.1 0.326 0.508 483 483 483 0.4 12.7 1.47 0.74 0.31 0.47 10 2
19 154 13 1.1 0.326 0.508 483 483 483 0.4 12.7 1.47 0.74 0.28 0.52 10
20 156 13 1.1 0.326 0.508 483 483 483 0.4 12.7 1.47 0.51 0.27 0.26 10
21 199 13 1.1 0.326 0.508 483 483 483 0.4 12.7 1.47 0.66 0.24 0.46 10
22 225 13 1.1 0.326 0.508 483 483 483 0.4 12.7 1.47 0.73 0.23 0.56 10
23 234 13 1.1 0.326 0.508 483 483 483 0.4 12.7 1.47 0.66 0.22 0.48 10
24 219 13 1.1 0.326 0.508 483 483 483 0.4 12.7 1.47 0.66 0.23 0.48 10
25 246 13 1.1 0.326 0.508 483 483 483 0.4 12.7 1.47 0.56 0.22 0.38 10
26 130 13 1.1 0.326 0.508 483 483 483 0.4 12.7 1.47 0.95 0.30 0.72 10
27 219 13 1.1 0.326 0.508 483 483 483 0.4 12.7 1.47 0.67 0.23 0.49 10
28 246 13 1.1 0.326 0.508 483 483 483 0.4 12.7 1.47 0.7 0.22 0.53 10
29 60 19 1.09 0.503 0.925 1099 819 610 0.63 25.4 1.33 1.8 0.42 1.38 5
30 100 19 1.09 0.503 0.925 1099 819 610 25.4 1.33 1.7 0.32 1.38 5
a
P / D = ⬁; single tube

Journal of Pressure Vessel Technology FEBRUARY 2011, Vol. 133 / 014002-3

Downloaded From: https://pressurevesseltech.asmedigitalcollection.asme.org on 06/18/2019 Terms of Use: http://www.asme.org/about-asme/terms-of-use

Table 2 Outline of references, tube configurations, and experimental considerations

sive theory for viscous damping of a cylinder in confined viscous Equation 共2兲 was verified against available data found in sev-
fluid. Rogers et al. 关25兴 developed a simplified version valid for eral references 关27,40,41,24,28兴. The agreement between the ex-
S ⬎ 3300 and D / De ⬍ 0.5, which covers most heat exchangers. periment and the theory is generally good, as shown in Fig. 4. The
Their simplified expression for ␨V 共in %兲 is experimental values are slightly lower than calculated for more

冋 冉 冊册
冦 冧
3
viscous liquids such as oils. The lower damping values measured
D in oils are probably due to the approximate nature of Eq. 共1兲 at
冉 冊冉 冊
1+
100␲ ␳D 2
2␯ 0.5
De lower Stokes Numbers, as shown in Fig. 5. However, this is not

冋 冉 冊册
␨␯ = 共2兲
冑8 m ␲ fD2 D 2 2 very significant from a practical point-of-view. Thus, we recom-
1− mend the use of Eq. 共2兲 to evaluate tube-to-fluid viscous damping.
De For every data point of Table 1, we calculated the viscous
where ␳ is the density of the fluid, m is the total mass of the tube damping ␨v and subtracted it from the overall damping ␨T. The
per unit length and includes the mass of fluid inside the tube mi, results, which represent damping at the supports ␨S are presented
the tube mass mt, and the hydrodynamic mass mh 共also called in Fig. 6. They include squeeze-film damping ␨SF and friction
added mass兲, and De is the effective diameter of the surrounding damping ␨F, thus
tubes or the inside diameter of the surrounding cylinder if the tube ␨T − ␨v = ␨S = ␨SF + ␨F 共6兲
is in an annulus.
The effective diameter is derived from data for a tube sur- Interestingly, the viscous damping between the tube and fluid is
rounded by rigid tubes. For a triangular tube bundle, Rogers et al. relatively small, and is generally between 10% and 30% of the
关25兴 found overall damping. Thus, most of the energy dissipations take place
De/D = 共0.96 + 0.50P/D兲P/D 共3兲
Similarly, for a square tube bundle, we found
De/D = 共1.07 + 0.56P/D兲P/D 共4兲
where P is the pitch between tubes. The hydrodynamic mass mh
= Cm共␲␳D2 / 4兲, where Cm is a coefficient to take into account the
effect of confinement, i.e.

冋 冉 冊册
D
1+
De
2

冋 冉 冊册
Cm = 2 共5兲
D
1−
De
The assumption of a cylinder surrounded by rigid cylinders was
made for simplicity. The dynamics of a cylinder surrounded by
flexible cylinders is much more complicated because the hydro-
dynamic coupling between cylinders makes many coupled modes
of vibration possible. Evaluation of the damping for each hydro-
dynamically coupled mode is beyond the state-of-the-art. At flow
velocities below that required for fluidelastic instability, several
modes of vibration are generally present at the same time. We
assume that the average damping of these modes is similar to the Fig. 4 Viscous damping data for a cylinder in confined †28‡
damping of a cylinder surrounded by rigid cylinders. Obviously, and unconfined liquids „all other data…: comparison between
this question requires further study. theory and experiment

014002-4 / Vol. 133, FEBRUARY 2011 Transactions of the ASME

Downloaded From: https://pressurevesseltech.asmedigitalcollection.asme.org on 06/18/2019 Terms of Use: http://www.asme.org/about-asme/terms-of-use

Fig. 7 Heat exchanger tube with N spans and „N − 1… interme-
diate supports

motions. Damping due to rocking motion should be less than that

due to lateral motion since it involves less relative motion at the
supports. Rocking motion is predominant in the lower modes,
which are generally more vulnerable to vibration problems. Thus,
rocking motion may be more important in practice.
The impacting category of dynamic interaction involves both
squeeze-film and impact dampings. As discussed in Sec. 2, we
Fig. 5 Viscous damping of cylinders in liquids versus Stokes
number
cannot at this stage treat damping due to impacting and the result-
ing traveling waves separately from the other damping mecha-
nisms.
Also, the energy dissipated by sliding may be due to metal-to-
at the support. We will see later on that friction damping appears metal contact friction forces or viscous shear forces or a combi-
to be less important. This means that squeeze-film damping is nation of the two, as discussed in Sec. 7. These considerations
dominant. apply to both lateral- and rocking-type motions.
The above finding is particularly important for crudded heat The total vibration energy in a tube is proportional to the num-
exchangers. If the clearance gap between tubes and tube-supports ber of spans N, assuming that span length and all other tube pa-
is solidly jammed with crud, support damping becomes very rameters are kept constant. The energy dissipated at the clearance
small. The tubes would then be more prone to excessive vibration supports by squeeze-film and friction dampings is obviously re-
and perhaps fatigue problems. lated to the number of clearance supports 共N − 1兲. In a two-span
heat exchanger tube with one support, there is the vibration energy
5 Damping at the Supports of two spans but only one support to dissipate energy. Thus,
The principal energy dissipation mechanisms at the supports damping should be less than for a tube with a large number of
共squeeze-film and friction dampings兲 obviously depend on the dy- spans and supports. Accordingly, the data was normalized, such
namic interaction between tubes and tube-supports, as discussed that
in Sec. 2. In real heat exchangers, the dynamic interaction be- ␨Sn = ␨SN/共N − 1兲 = 共␨T − ␨V兲N/共N − 1兲 共7兲
tween tubes and supports is generally a combination of sliding,
impacting, and scuffing, and involves both lateral and rocking where ␨Sn is the normalized damping related to the supports. Fig-
ure 6 shows ␨Sn as a function of frequency.
Some energy may be dissipated due to friction at the tube-to-
tube sheet joints. This will obviously depend on the type of tube
joint. For a tube carefully clamped at both ends but without inter-
mediate support, we have measured damping ratios of roughly
0.07% over a frequency range of 40–150 Hz 关15兴. Thus, friction in
the tube joints may contribute some damping, depending on the
type of joints. Since it is relatively small, it is prudent not to
consider it unless it is reasonably well known for the heat ex-
changer design being considered.

6 Squeeze-Film Damping
Squeeze-film damping is due to the fluid forces in the annular
clearance between the tube and its supports. Fluid forces are gen-
erated as the tube motion squeezes the fluid between the tube and
tube-support. This phenomenon was first studied by Mulcahy 关23兴
for a cylinder vibrating within a concentric cylinder of finite
length. Chen 关21兴 and Jendrzejczyk 关24兴 extended this work to the
case of a two-span heat exchanger tube with one intermediate
support. We shall extend this work to a multispan tube to develop
an equation for squeeze-film damping. This expression will be
subsequently reviewed in terms of relevance, accuracy, and appli-
cability. Several assumptions and linearizations will be required as
explained later.
The equation of motion of a multispan heat exchanger tube with
Fig. 6 Damping due to tube-supports in multispan heat ex- N spans and N − 1 intermediate concentric tube-supports 共see Fig.
changer tubes 7兲 may be expressed by

Journal of Pressure Vessel Technology FEBRUARY 2011, Vol. 133 / 014002-5

Downloaded From: https://pressurevesseltech.asmedigitalcollection.asme.org on 06/18/2019 Terms of Use: http://www.asme.org/about-asme/terms-of-use

⳵ 4u
EI + 关CV + C̄SF␦共x1兲 + C̄SF␦共x2兲 + . . . . + C̄SF␦共xJ兲 + . . . .
⳵ x4
⳵u ⳵ 2u
+ C̄SF␦共xN−1兲兴 +m 2 =0 共8兲
⳵t ⳵t
where EI is the flexural rigidity, u is the tube vibration response at
any point x along the tube and at any time t, x1, x2, xJ, and xN−1
are the locations of the first, second, Jth, and last tube-supports,
CV is the tube-to-fluid viscous damping coefficient, C̄SF is the
integration of CSF over the thickness L of the support, CSF is the
squeeze-film damping coefficient, and ␦共xJ兲 is a delta function
corresponding to the location of each support. This is required to
relate the energy dissipated locally by squeeze-film damping at the
Jth tube-support to the vibrating energy in the whole tube.
It is convenient to express the vibration response in terms of
generalized coordinates qi共t兲 such that
n

u共x,t兲 = 兺 q 共t兲␾ 共x兲

i=1
i i 共9兲

where ␾i共x兲 is the mass normalized mode shape of the ith mode
defined as

冕
ᐉ
m␾i2共x兲dx = 1 共10兲 Fig. 8 Linearization of three-dimensional factor K
0

It may be shown from Eqs. 共8兲 and 共9兲 that

冉 冊
冤 冥
2z

冉 冊
d 2q i dqi cosh
+ 2 ␨ i␻ i + ␻ i2q i = 0 共11兲 ␲ D Ds

冉 冊
dt2 dt CSF = ␳D ␻i关− Im共H兲兴 1 − 共15兲
4 Ds − D L
where cosh
Ds
N−1
Thus,
兺
C␯ 1

冕 冉 冊
␨i = + C̄SF␸Ji
2
共12兲 +L/2
2m␻i 2␻i J=1 ␲ D
C̄SF = CSFdz = L ␳ D 2K ␻i关− Im共H兲兴 共16兲
4 Ds − D
since −L/2

where z is a point along the width of the support, Ds is the support

␨i = ␨vi + ␨SFi hole diameter, and K is a factor to take into account three-
dimensional flow effects 共i.e., side-leakage due to the finite length
N−1
of the support兲. The factor K may be expressed as
兺
1

冉冊
␨SFi = C̄SF␾Ji
2
共13兲
2␻i J=1 D L
K=1− tanh 共17兲
L D
where ␾Ji represents the ith normalized modal amplitude at the
Jth support location. As shown in Fig. 8, K may be linearized as
It is convenient to define a characteristic tube length ᐉm, which K = 0.225共L/D兲1.3 共18兲
we take as the average of the three longest spans. This is based on
the assumption that the lower modes, which involve primarily the for L / D between 0.5 and 2.0, which is a realistic range for heat
longer spans, dominate the vibration response. Modes involving exchanger tube-supports.
the longest spans tend to be more vulnerable because the ratio of The function Im共H兲 derived by Mulcahy 关23兴 is very compli-
energy dissipated at the support over the vibration energy in the cated. However, it is practically linear for values of 冑8v / ␻ / 共Ds
tube is lower. Also, the relative motion at the support tends to be − D兲 greater than 0.3, as shown in Fig. 9, such that
more of the rocking-type than the lateral-type. When higher
modes and shorter spans govern the vibration response, then, the − Im共H兲 = 共48␯/␻兲/共Ds − D兲2 共19兲
characteristic span length should be based on these shorter spans. where ␻ = 2␲ f. For many realistic heat exchangers, 冑8v / ␻ / 共Ds
This could happen when there are high flow velocities locally
− D兲 is near or greater than 0.3; thus, the linearization formulated
in Eq. 共19兲 is acceptable. For 冑8v / ␻ / 共Ds − D兲 smaller than 0.3,
such as in entrance or exit regions. This is a crude attempt at
weighting the effect of mode shape in the squeeze-film damping
predictions. Thus, for a given mode Eq. 共19兲 would yield conservative damping predictions.
Now Eqs. 共14兲, 共16兲, 共18兲, and 共19兲 are inserted into Eq. 共13兲 to
N−1 obtain
Aiⴱ
兺␾ 共14兲
冉 冊冉 冊 冋 册 冋 册
2
=
Ji
mᐉm ␲␳D2 D AiⴱL L 1.3
48v
J=1 ␨SFi = 0.225 ␻i
8m␻i Ds − D mᐉm D ␻i共Ds − D兲2
where Aⴱi
is simply a factor integrating the effect of the relative
motion between tube and support for all the supports. 共20兲
From Refs. 关21,24兴 Rearranging

014002-6 / Vol. 133, FEBRUARY 2011 Transactions of the ASME

Downloaded From: https://pressurevesseltech.asmedigitalcollection.asme.org on 06/18/2019 Terms of Use: http://www.asme.org/about-asme/terms-of-use

consider both metal-to-metal contact friction and viscous shear
sliding forces.
Viscous shear forces are related to viscosity ␮ = ␳␯, relative
velocity U = ⳵u / ⳵t, bearing length b, and liquid film height h.
Since there is no hydrodynamic lubrication, the film height is
governed by surface asperities and, therefore, is not a variable
parameter but a constant. The viscous shear force per unit support
thickness FVSs is a function of ␳v ⳵ u / ⳵t. For a tube inside a sup-
port, the bearing length is not easily defined. However, it is related
to the tube diameter and to the inverse of the tube-to-support
clearance, thus, b = ␺关D / 共Ds − D兲兴. Similar to Sec. 6, we define a
viscous shear damping coefficient CVS, such that
FVS = CVS ⳵ u/⳵ t = ␺关D/共Ds − D兲兴␳v ⳵ u/⳵ t 共23兲
from which,
CVS = v␳␺关D/共Ds − D兲兴 共24兲
and the integrated coefficient C̄VS

冕
+L/2
C̄VS = CVSdz 共25兲
−L/2

Thus

C̄VS = v␳L␺关D/共Ds − D兲兴␺共L/D兲 共26兲

where the function of L / D is added to take into account three-
dimensional effects or side flow. The corresponding viscous shear
damping ratio ␨VSi for multisupport heat exchanger tube is simi-
Fig. 9 Damping and hydrodynamic mass functions Im„H… and
Re„H… †23‡
larly expressed by
N−1
C̄VSAiⴱ
兺
1
␨VSi = C̄VS␾Ji
2
= 共27兲

冉 冊冉 冊冉 冊 冉 冊冋 册 2␻i J=1 2m␻iᐉm

␳D2 L L 1.3
D v
␨SFi = Ai 共21兲
m ᐉm D Ds − D f共Ds − D兲2 where Ai is a factor integrating the effect of the relative sliding
motion between the tube and support for all the supports.
We now have to estimate constant Ai and validate the form of Eq. Inserting Eqs. 共14兲 and 共26兲 in Eq. 共27兲, and rearranging, we get
共21兲 with the experimental data. This will be discussed in Sec. 8.

7 Damping Due to Sliding ␨VSi = Ai⬙ 冉 冊冉 冊 冉

v␳
mf
L
ᐉm
␺
D
DS − D
␺
L
D
冊冉 冊 共28兲
Sliding of the tube within the support may generate metal-to- The form of Eq. 共28兲 is similar to Eq. 共21兲. However, we have no
metal contact friction and viscous shear forces. If the conditions way of separating viscous shear damping from squeeze-film
are such that hydrodynamic lubrication occurs, the tube becomes damping in the data of Fig. 6. Since their formulation is similar,
supported entirely by a film of liquid and metal-to-metal contact is we shall consider them together as squeeze-film damping and as-
prevented. Under these conditions, the liquid film between the sume they both follow the same formulation.
tube and the support is thicker than the surface asperities of the Damping of heat exchanger tubes in gases due to metal-to-
tube and the support. The problem may be studied by using hy- metal contact friction forces is discussed in Ref. 关15兴. Friction
drodynamic lubrication theory. According to Ouziaux and Perrier damping appears independent of most heat exchanger parameters
关42兴 the thickness of the liquid film h is expressed in dimension- except for support width and for the number of supports when
less form as there are very few supports. There is no reason to believe that
h/b = K p冑␳vU/p 共22兲 metal-to-metal friction damping should depend on different pa-
rameters in liquids.
where b is the effective bearing length and K P is a dimensionless
coefficient, having a value of roughly 0.40 for hydrodynamic lu-
brication, U is the relative velocity between the two load bearing
surfaces, and p is the bearing load per unit width. For a heat 8 Semi-Empirical Formulation of Tube-Support
exchanger tube within a support, b is much less than D, and U can Damping
be taken as ␻u. For a typical tube of D = 20 mm, lm = 1.0 m, m The problem now is to compare the damping formulation of
= 1.0 kg/ m, ␻ = 200 rad/ s, vibrating at an amplitude u Eqs. 共21兲 and 共28兲 to the data of Fig. 6. The objective is to derive
= 100 ␮m within a support of thickness L = 10 mm, in water of semi-empirical expressions to formulate tube damping in liquids.
␯ = 10−6 m2 / s and ␳ = 1000 kg/ m3, U becomes 0.02 m/s and p Several assumptions and simplifications are necessary since the
= 1000 N / m, assuming that the loading at the support is roughly available data are not sufficient to study all the relevant param-
equal to the tube weight. We calculate h / b = 57⫻ 10−6. Since b eters. Furthermore, the large scatter in the data makes this task
Ⰶ D = 20 mm, h Ⰶ 1.14 ␮m. For that tube there should be no hy- very difficult. Therefore, we shall proceed in stages. Starting from
drodynamic lubrication since the tube and support surface asperi- the simplest formulation, we shall develop a more physically
ties are larger than 1.0 ␮m. Thus, it appears that for many heat based expression.
exchanger tubes in liquids, the conditions for hydrodynamic lubri- The data of Fig. 6 show that damping at the supports decreases
cation are not achieved except perhaps for higher viscosity liquids with frequency. Since the theory of Eq. 共21兲 also implies a de-
such as oils or for very light support loadings, which could be crease in squeeze-film damping proportional to the inverse of fre-
encountered in vertical heat exchangers. Therefore, we have to quency, we accept that this damping mechanism follows the rela-

Journal of Pressure Vessel Technology FEBRUARY 2011, Vol. 133 / 014002-7

Downloaded From: https://pressurevesseltech.asmedigitalcollection.asme.org on 06/18/2019 Terms of Use: http://www.asme.org/about-asme/terms-of-use

Fig. 11 Effect of support thickness parameter „L / lm… on damp-
ing due to tube-supports

conventional heat exchangers. Of course this trend would not nec-

essarily apply for very tight clearances such as the case of crudded
supports or for very large clearances.
Fig. 10 Squeeze-film damping of multispan heat exchanger Consequently, the parameter D / 共Ds − D兲 was removed from the
tube in water formulation. This also indicated that the term 共Ds − D兲 in the pa-
rameter ␯ / 关f共Ds − D兲2兴 is not correct. Thus, 共Ds − D兲 was elimi-
nated since we do not, as of yet, know of a better term. Clearly, a
tionship ␨SF⬁1 / f. On the other hand, the component of support more appropriate term must be found to represent confinement in
damping due to metal-to-metal friction is considered independent the dimensionless parameter ␯ / 关f共Ds − D兲2兴, which is the equiva-
of frequency, as discussed in Ref. 关15兴. lent of the Stokes number.
We know that tube-support damping ␨s comprises two damping The effect of viscosity could not easily be studied since most of
mechanisms, i.e. the data were for water at room temperature. Taylor et al. 关19兴
studied damping of multispan tubes at 25° C, 60° C, and 90° C
␨s = ␨SF + ␨F or ␨sn = ␨SFn + ␨Fn 共29兲 and found little effect of temperature. Nagpal et al. 关37兴 also mea-
As a first attempt to correlate the data of Fig. 6, we simply sub- sured damping at 2 ° C and 65° C. Although there was a lot of
tracted a constant value for friction damping ␨Fn so that the re- scatter in the data, they found surprisingly little effect of tempera-
maining damping satisfies the relationship ␨SF⬁1 / f. This required ture on damping. The effect of viscosity on squeeze-film damping
a friction damping value of roughly 0.2%. The remaining squeeze- obviously needs to be studied further.
film damping ␨SFn satisfies the relationship ␨sn = 50/ f, as shown in We did not see much effect of the parameter 共L / D兲, indicating
Fig. 10. Thus, we get the following expression for tube-support that side-leakage or three-dimensional effects are not significant in
damping, in percent the data studied. Thus, we are left with a simplified expression for
squeeze-film damping ␨SFn, that is
␨sn = 50/f + 0.2
冉 冊冉 冊冉 冊
共30兲
1600 ␳D2 L 0.5
To obtain conservative but realistic damping guidelines, we took ␨SFn = 共31兲
the damping level corresponding to the lower decile of the avail- f m ᐉm
able damping data. This means that 90% of the data is above the This expression is a reasonable fit of the minimum damping data
suggested relationship and 10% below. This is called the mini- available, as shown in Fig. 12.
mum realistic damping level. However, it is quite different from the theoretical expression of
As a second step, we attempted to develop a more physically Eq. 共21兲. The most likely explanation is that the theory was de-
based model for squeeze-film damping. We started by assuming veloped for a tube concentric within its support 关23兴, whereas in
that friction damping is negligible since it is relatively small as practice, the tube is very likely to touch the support on one side.
shown above. Then, we correlated systematically the parameters The hydrodynamic forces, and hence the damping forces, are
of Eq. 共21兲 with the data of Fig. 6 to obtain the best possible fit. much larger near the contact area. This phenomenon is controlled
The dimensionless thickness L / ᐉm appears to be an important by the local proximity h of the tube to the support not by the
parameter, as discussed by Hartlen 关20兴. It is also shown in Fig. average clearance Ds − D between tube and support. This explains
11. Although there is a lot of scatter in the data, it appears that why the parameter D / 共Ds − D兲 is not significant. Since the tube is
L / ᐉm exponent 0.5 best fits the data. The parameter 共␳D2 / m兲 was generally very near the support, the ratio L / h is large, the imped-
not studied systematically since it only varied over a narrow ance to side flow is large and three-dimensional effects are mini-
range. However, the data appeared to follow the theory, and the mized. Thus, the effect of the parameter 共L / D兲 should be small.
parameter 共␳D2 / m兲 was accepted as correct. As outlined in Table 2, there is some albeit limited data for
We also studied the effect of the parameter D / 共Ds − D兲. Al- support types other than drilled holes. Most of this data are for
though Hartlen 关20兴 and Shin et al. 关26兴 observed an increase in vertical tubes with supports such as broached holes and egg
damping with D / 共Ds − D兲, we found no significant trend over the crates. These data are related to nuclear steam generator studies.
range of the available data. Most of the data were for diametrical Except for one data point, all the other data are for horizontal
clearances between 0.40 mm and 0.80 mm, which is normal for tubes with drilled-hole-type supports. Interestingly, the vertical

014002-8 / Vol. 133, FEBRUARY 2011 Transactions of the ASME

Downloaded From: https://pressurevesseltech.asmedigitalcollection.asme.org on 06/18/2019 Terms of Use: http://www.asme.org/about-asme/terms-of-use

Fig. 14 Comparison between the tube-support damping model
„squeeze-film and friction… and the experimental data
Fig. 12 Comparison between tube-support damping param-
eter and experimental data for heat exchanger tube damping in
water „squeeze-film model only… Combining Eq. 共32兲 for friction damping and Eq. 共31兲 for
squeeze-film damping, we get a complete formulation for support
damping 共i.e., ␨sn = ␨SFn + ␨Fn兲:
tube data with other support types is not singularly different than
the horizontal tube 共drilled-hole兲 data. This also suggests that
␨sn =
f
冉 冊冉 冊
1460 ␳D2
m
L
ᐉm
0.5
冉 冊
+ 0.5
L
ᐉm
0.5
共33兲
damping is controlled by local contact and not by the overall
as shown in Fig. 14. This model is more physically based since it
geometry of the support. In this respect the local contact of a tube
contains most of the important parameters, which influence heat
on a lattice bar is not significantly different than the local contact
exchanger tube damping.
of a tube with a drilled hole.
In reality, the tube will be at a slight angle with the axis of the
support most of the time, particularly for rocking-type motion.
9 Discussion
The tube and the axis of the support are parallel only a small The current situation is far from ideal, as illustrated in Figs. 10,
portion of the time. Thus, the contact between the tube and sup- 12, and 14. There is a large scatter in the data and the trends are
port takes place on a point or a corner rather than on a line, as not always clear. This is due to the rather variable quality of the
shown in Fig. 13. This would tend to reduce the effect of support data and, more likely, to the very nature of the problem. Heat
width or the dimensionless parameter L / ᐉm. It may explain the exchanger tube damping depends on both deterministic and statis-
less than unity exponent, that is, 0.5, for the parameter L / ᐉm. tical parameters. Support thickness, clearance between the tube
The latter model, formulated by Eq. 共31兲, neglects the contri- and support, tube diameter, spacing between supports, and tube
bution of friction damping. This tends to slightly underestimate frequency are deterministic parameters. Their effects should be
damping at higher tube frequencies, as may be seen in Fig. 12. predictable. On the other hand, many parameters are difficult to
Thus, as a third step, we shall add a friction damping term of the control such as support alignment, tube straightness, side loads,
following form: and relative position and motion of the tube within the support.
These parameters are statistical in nature and probably contribute
␨Fn = B共L/ᐉm兲0.5 共32兲 to much of the scatter in the damping data. It explains why heat
as suggested by Pettigrew et al. 关15兴. exchanger tube damping has been so elusive thus far.
Further damping measurements on many realistic heat ex-
changer tube configurations are required to provide the statistical
basis for a validated design criterion. However, we cannot wait
until all results are available. “Best-effort” design recommenda-
tions are needed now.

10 Design Recommendations
A conservative but realistic criterion for damping of heat ex-
changer tubes in liquids is required. As already mentioned, this is
achieved by taking damping values at roughly the lower decile of
the available damping data. This is a reasonably conservative ap-
proach and yet it is practical from a design point-of-view. It is not
unduly constrained by unusually low values. Based on Figs. 3, 10,
12, and 14, several approaches are possible from the simplest
Fig. 13 Type of contact between the tube and the support criterion based directly on the available data to more physically

Journal of Pressure Vessel Technology FEBRUARY 2011, Vol. 133 / 014002-9

Downloaded From: https://pressurevesseltech.asmedigitalcollection.asme.org on 06/18/2019 Terms of Use: http://www.asme.org/about-asme/terms-of-use

based expressions, taking into account the important damping
mechanisms and the relevant tube and support parameters. 冉 冊冉 冊
␨F = 0.5
5
6
0.015
1.0
1/2
= 0.05%
10.1 Simple Criterion Based on Available Data. A simple Thus, the total tube damping ␨T is
criterion can be derived directly from the available experimental
data of Fig. 3. This would yield a damping ratio in percent, ex- ␨T = ␨V + ␨SF + ␨F
pressed by
␨T = 0.47 + 1.19 + 0.05 = 1.71%
␨ = 60/f for f ⬍ 50 Hz
This is a realistic damping value for a heat exchanger tube in
␨ = 8.5/冑 f for 50 ⬍ f ⬍ 200 Hz water.

␨ = 0.6% for f ⬎ 200 Hz 共34兲 12 Concluding Remarks

This criterion takes into account the effect of frequency, as shown There are three important energy dissipation mechanisms that
in Fig. 3. It is probably adequate for many normal heat exchang- contribute to damping of heat exchanger tubes in liquids. These
ers. are viscous damping between the tube and liquid, squeeze-film
damping in the clearance between the tube and support, and fric-
tion damping at the support.
10.2 Criterion Based on the Formulation of Three Energy Viscous damping only accounts for approximately 25% of the
Dissipation Mechanisms. A better criterion is one based on Eqs. total damping of a typical tube. Thus, about 75% of the damping
共2兲 and 共33兲. It takes into account the three most important energy energy is dissipated at the support. Squeeze-film damping appears
dissipation mechanisms, that is, viscous damping ␨V, squeeze-film to be the most important energy dissipation mechanism.
damping ␨SF, and friction damping ␨F. Thus The available expression to calculate viscous damping was sim-
␨T = ␨V + ␨SF + ␨F 共35兲 plified and validated against experimental data. Squeeze-film
damping is related to support width and is inversely proportional
Substituting Eqs. 共2兲 and 共33兲 in Eq. 共35兲, we get to tube frequency. Semi-empirical expressions have been devel-

冋 冉 冊 册 冉 冊冉 冊 冉 冊
冦 冧
D 3 oped to formulate squeeze-film and friction dampings.
1+ An expression 共Eq. 共36兲兲 comprising viscous, squeeze-film, and
100␲ De ␳D2 2v 1/2
N−1

冋 冉 冊册
␵T = + friction dampings is recommended for design purposes.
冑 8 D 2 2
m ␲ fD2 N
1−
De Acknowledgment

⫻冋 冉 冊冉 冊 冉 冊 册
共1460兲 ␳D
f m
2
L
ᐉm
1/2
+ 0.5
L
ᐉm
1/2
共36兲
The authors have benefited from discussions with B. Villard of
Centre d’Études Nucléaires de Saclay; H.G.D. Goyder of UKAEA
Harwell; R.T. Hartlen of Ontario Hydro; and C.E. Taylor and J.H.
where the support damping mechanisms are adjusted by 共N Tromp of Chalk River Laboratories. Their contributions are grate-
− 1兲 / N to take into account the ratio of number of supports over fully acknowledged. Z.L. Qiao of Xian Jiaotong University,
number of spans. China, and J. Albrecht of Chalk River Laboratories contributed to
Although somewhat speculative, Eq. 共36兲 formulates all impor- the analysis of the data.
tant energy dissipation mechanisms and fits the data best. Thus, it
is our current recommendation as a damping criterion for design
Nomenclature
purposes. However, if the damping ratio predicted by this equa-
tion is less than 0.6%, we recommend taking a minimum value of A ⫽ modal factor
0.6%. As shown on Fig. 3, a minimum damping of 0.6% appears b ⫽ effective bearing length
reasonable. Cm ⫽ confinement coefficient
C ⫽ damping coefficient
C̄ ⫽ damping coefficient integrated over support
thickness
11 Sample Calculations
D ⫽ tube diameter 共m兲
To illustrate the application of Eq. 共36兲, we take a typical mul- De ⫽ equivalent diameter of surrounding tubes 共m兲
tispan heat exchanger tube of diameter D = 20 mm, mass includ- Ds ⫽ support diameter
ing hydrodynamic mass m = 1.0 kg/ m, and frequency f = 50 Hz, EI ⫽ flexural rigidity
within a triangular tube bundle of P / D = 1.3, supported by N − 1 f ⫽ frequency 共Hz兲
= 5 supports of thickness L = 15 mm, of characteristic span length h ⫽ film thickness
ᐉm = 1.0 m, operating in water of viscosity ␯ = 10−6 m2 / s and of K ⫽ side-leakage factor
density ␳ = 1000 kg/ m3 Kp ⫽ dimensionless hydrodynamic lubrication
De/D = 共0.96 + 0.50P/D兲P/D = 2.09 coefficient
L ⫽ support thickness 共m兲
Equation 共36兲 gives ᐉm ⫽ characteristic span length 共m兲

冉 冊
冦 冧
1 3 m ⫽ total mass per unit length 共mt + mh + mi兲 共kg/m兲

冋 冉 冊册 冉 冊
100␲
1+
2.09 1000 ⫻ 0.022 N ⫽ number of tube spans
␨v = p ⫽ bearing load per unit width
冑 8 1 2 2
1.0 q ⫽ generalized coordinates
1−
2.09 P ⫽ tube pitch 共m兲

冉 冊
S ⫽ Stokes number
2 ⫻ 10−6 1/2
⫻ = 0.47% t ⫽ time
␲ ⫻ 50 ⫻ 0.022 u ⫽ vibration amplitude

冉 冊冋冉 冊冉 冊冉 冊 册
U ⫽ sliding velocity
5 1460 1000 ⫻ 0.022 0.015 1/2
x ⫽ distance along tube
␨SF = = 1.19%
6 50 1.0 1.0 xj ⫽ support location

014002-10 / Vol. 133, FEBRUARY 2011 Transactions of the ASME

Downloaded From: https://pressurevesseltech.asmedigitalcollection.asme.org on 06/18/2019 Terms of Use: http://www.asme.org/about-asme/terms-of-use

z ⫽ position along support 429.
关15兴 Pettigrew, M. J., Goyder, H. G. D., Qiao, Z. L., and Axisa, F., 1986, “Damping
␨ ⫽ damping ratio 共%兲 of Multispan Heat Exchanger Tubes—Part 1: In Gases,” Proceedings of the
␮ ⫽ dynamic viscosity 共kg/ m s兲 Symposium on Special Topics of Structural Vibration, ASME Pressure Vessels
␯ ⫽ kinematic viscosity 共m2 / s兲 and Piping Conference, Chicago, IL.
关16兴 Pettigrew, M. J., Rogers, R. J., and Axisa, F., 1986, “Damping of Multispan
␳ ⫽ fluid density 共kg/ m3兲 Heat Exchanger Tubes—Part 2: In Liquids,” Proceedings of the Symposium on
␴ ⫽ surface tension 共N/m兲 Special Topics of Structural Vibration, ASME Pressure Vessels and Piping
␦共 兲 ⫽ delta function Conference, Chicago, IL.
关17兴 Pettigrew, M. J., Rogers, R. J., and Taylor, C. E., 1985, “Damping of Heat
␾␫共x兲 ⫽ normalized mode shape Exchanger Tubes,” Atomic Energy of Canada Limited, Report No. AECL-
␻ ⫽ angular frequency, 2␲ f 8701.
␺关 兴 ⫽ function of 关18兴 Pettigrew, M. J., Axisa, F., and Qiao, Z. L., 1985, “Amortissement des Tubes
d’Échangeur de Chaleur en Milieu Liquide,” Commissariat à 1’Énergie Atom-
Subscript ique, CEN Saclay, France, Report No. DEMT/85/462.
F ⫽ friction 关19兴 Taylor, C. E., Pettigrew, M. J., Dickinson, T. J., Currie, I. G., and Vidalou, P.,
1998, “Vibration Damping in Multispan Heat Exchanger Tubes,” ASME J.
h ⫽ hydrodynamic Pressure Vessel Technol., 120共3兲, pp.283–289.
i ⫽ mode order 关20兴 Hartlen, R. T., 1974, “Effect of Support-Point Details Upon Natural Frequency
J ⫽ support order and Damping of Heat Exchanger Tubes: Preliminary Laboratory Investiga-
tion,” Ontario Hydro Research, Report No. 74-440-K.
l ⫽ liquid 关21兴 Chen, S. S., 1983, “Design Guide for Calculating Fluid Damping for Circular
n ⫽ normalized Cylindrical Structures,” Argonne National Laboratory, Report No. ANL-83-54.
p ⫽ pitch 关22兴 Collinson, A. E., and Warneford, I. P., 1978, “Vibration Tests of Single Heat
s ⫽ support Exchanger Tubes in Air and Static Water,” Proceedings of the Second Inter-
national Conference on Vibration in Nuclear Plant, Keswick, UK, Paper No.
SF ⫽ squeeze-film 3.3.
T ⫽ total 关23兴 Mulcahy, T. M., 1980, “Fluid Forces on Rods Vibrating in Finite Length An-
t ⫽ tube or cylinder nular Regions,” ASME J. Appl. Mech., 47共2兲, pp. 234–240.
V ⫽ viscous 关24兴 Jendrzejczyk, J. A., 1984, “Dynamic Characteristics of Heat Exchanger Tubes
Vibrating in a Tube Support Plate Inactive Mode,” Argonne National Labora-
tory, Report No. ANL-84-39.
References 关25兴 Rogers, R. J., Taylor, C., and Pettigrew, M. J., 1984, “Fluid Effects on Multi-
span Heat Exchanger Tube Vibration,” ASME Paper No. H00316.
关1兴 Goyder, H. G. D., 1982, “Measurement of the Natural Frequencies and Damp-
关26兴 Shin, Y. S., Jendrzejczyk, J. A., and Wambsganss, M. W., 1977, “Vibration of
ing of Loosely Supported Tubes in Heat Exchangers,” Proceedings of the
a Heat Exchanger Tube With Tube/Support Impact,” ASME Paper No. 77-
Third International Conference on Vibration in Nuclear Plants, Keswick, UK,
JPGC-NE-5.
Paper No. 2.7.
关27兴 Kim, B. S., Pettigrew, M. J., and Tromp, J. H., 1988, “Vibration Damping of
关2兴 Haslinger, K. H., and Thompson, J. P., 1980, “Determination of Damping
Heat Exchanger Tubes in Liquids: Effects of Support Parameters,” J. Fluids
Properties in Lightly Damped Systems,” Proceedings of the ASME-WAM
Struct., 2, pp. 593–614.
HTD, Chicago, IL, Vol. 9, pp. 53–63.
关28兴 Chen, S. S., Wambsganss, M. W., and Jendrzejczyk, J. A., 1976, “Added Mass
关3兴 Chen, S. S., and Jendrzejczyk, J. A., 1980, “Flow-Velocity-Dependence of
and Damping of a Vibrating Rod in Confined Viscous Fluids,” ASME J. Appl.
Damping in Tube Arrays Subjected to Liquid Cross Flow,” ASME Paper No.
Mech., 43, pp. 325–329.
80-C2/PVP-129.
关4兴 Scriven, A. H., and Hopley, C. E., 1980, “The Effect of an Internal Two-Phase 关29兴 Rogers, R. J., and Ahn, K. J., 1986, “Fluid Damping and Hydrodynamic Mass
Flow on the Damping Characteristics of a Vibrating Tube,” Proceedings of the in Finite Length Cylindrical Squeeze Films With Rectilinear Motion,” Flow-
European Two-Phase Flow Group Meeting, Strathclyde University, Glasgow, Induced Vibration,” ASME, New York, Vol. 104, pp. 99–105.
UK, Paper No. F2. 关30兴 Aikin, J. A.,1977, personal communication.
关5兴 Gravelle, A., Ross, A., Pettigrew, M. J., and Mureithi, N. W., 2007, “Damping 关31兴 Antunes, J., Villard, B., and Axisa, F., 1985, “Cross-Flow Induced Vibration of
of Tubes Due to Internal Two-Phase Flow,” J. Fluids Struct., 23共3兲, pp. 447– U-Bend Tubes of Steam Generator,” Proceedings of the Eighth International
462. Conference on Structural Mechanics in Reactor Technology, Brussels, Bel-
关6兴 Carlucci, L. N., and Brown, J. D., 1983, “Experimental Studies of Damping gium, Paper No. F16/9.
and Hydrodynamic Mass of a Cylinder in Confined Two-Phase Flow,” ASME 关32兴 Gorman, D. J., 1982, personal communication.
J. Vib., Acoust., Stress, Reliab. Des., 105, pp. 83–89. 关33兴 Gorman, D. J., 1983, “Experimental Studies of the Liquid Cross-Flow Induced
关7兴 Pettigrew, M. J., Tromp, J. H., and Mastorakos, J., 1985, “Vibration of Tube Vibration of Multispan Heat Exchanger Tube Bundles in Triangular and
Bundles Subjected to Two-Phase Cross Flow,” ASME J. Pressure Vessel Tech- Square Arrays,” Proceedings of the 20th Congress of the IAHR, Moscow, Vol.
nol., 107, pp. 335–343. IV, pp. 415–426.
关8兴 Pettigrew, M. J., Taylor, C. E., and Kim, B. S., 1989, “Vibration of Tube 关34兴 Halle, H., and Wambsganss, M. W., 1980, “Tube Vibration in Industrial Size
Bundles in Two-Phase Cross Flow—Part 1: Hydrodynamic Mass and Damp- Heat Exchanger,” Argonne National Laboratory, Report No. ANL-CT-80-18.
ing,” ASME J. Pressure Vessel Technol., 111, pp. 466–477. 关35兴 Heilker, W. K., 1978, personal communication.
关9兴 Pettigrew, M. J., and Taylor, C. E., 2004, “Damping of Heat Exchanger Tubes 关36兴 Kobate, K., Suzuki, M., Koishikawa, A., and Hashimoto, T., 1980, “Vibration
in Two-Phase Flow: Review and Design Guidelines,” ASME J. Pressure Vessel Testing of a Straight Tube Type Steam Generator for FBR,” ASME Publica-
Technol., 126共4兲, pp. 523–533. tion, PVP-41, pp. 19–31.
关10兴 Fisher, N. J., Han, Y., Guérout, F. M., and Janzen, V. P., 2005, “Comparison of 关37兴 Nagpal, V., Campagna, A. O., and Pettigrew, M. J., 1981, “Tube Vibration in
Predicted and Observed Fretting-Wear Damage in Nuclear Steam Generators,” the Entrance Region of a Steam Generator for a 600 MW Nuclear Power
ASME Paper No. PVP2005-71390. Station,” Atomic Energy of Canada Limited, Report No. CRNL-2093.
关11兴 Morandin, G. D., and Sauvé, R. G., 2003, “Probabilistic Assessment of Fret- 关38兴 Pettigrew, M. J., Platten, J. L., and Sylvestre, Y., 1973, “Experimental Studies
ting Wear in Steam Generator Tubes Under Flow-Induced Vibrations,” ASME on Flow Induced Vibration to Support Steam Generator Design, Part II: Tube
Paper No. PVP2003-2081. Vibration Induced by Liquid Cross-Flow in the Entrance Region of a Steam
关12兴 Rao, M. S. M., Laskowski, L. J., Srikantiah, G. S., and Ahluwalia, K. S., 1997, Generator,” Proceedings of the International Symposium on Vibration Prob-
“Prediction of Tube Wear Due to Flow-Induced Vibration in PWR Steam Gen- lems in Industry, Keswick, UK, Paper No. 424. Also, Atomic Energy of
erators,” Proceedings of the Fourth International Symposium on Fluid- Canada Limited Report No. AECL-4515.
Structure Interaction, Aeroelasticity, Flow-Induced Vibration and Noise, 关39兴 Wambsganss, M. W., Halle, H., and Lawrence, W. P., 1981, “Tube Vibration in
ASME Congress and Exhibition, Dallas, TX, Nov. 16–21, Vol. 53, pp. 257– Industrial Size Test Heat Exchanger 共30° Triangular Layout—Six Crosspass
264. Configuration兲,” Argonne National Laboratory, Report No. ANL-CT-81-42.
关13兴 Pettigrew, M. J., Carlucci, L. N., Ko, P. L., Holloway, A. G. L., and Campa- 关40兴 Skop, R. A., Ramberg, S. E., and Ferer, K. M., 1976, “Added Mass and
gna, A. O., 1985, “Computer Techniques to Analyze Shell-and-Tube Heat Damping Forces on Circular Cylinders,” ASME Paper No. 76-Pet-3.
Exchangers,” Heat Transfer Eng., 6共3兲, pp. 40–52. 关41兴 Rogers, R. J., 1985, personal communication.
关14兴 Gorman, D. J., 1988, “Exact Analytical Solutions for the Free Vibration of 关42兴 Ouziaux, R., and Perrier, J., 1966, “Mécanique des Fluides Appliquée,” Flu-
Steam Generator U-Tubes,” ASME J. Pressure Vessel Technol., 110, pp. 422– ides Incompressibles, 2nd ed., Dunod, Paris, Vol. 1.

Journal of Pressure Vessel Technology FEBRUARY 2011, Vol. 133 / 014002-11

Downloaded From: https://pressurevesseltech.asmedigitalcollection.asme.org on 06/18/2019 Terms of Use: http://www.asme.org/about-asme/terms-of-use