UCRL-52342

EFFECTIVE MASS AND DAMPING OF

SUBMERGED STRUCTURES
R. G. Dong
April J, J97S

Mhsm
Work performed under the auspices of the U.S. Department of
Energy by the UCLLL under contract number W-7405-ENG-48.

LAWRENCE
LIVERMORE
LABORATORY
University otCaitomtaAJvermore

DISTRIBUTION OF THIS n n r , , ^
_,., v-f i tub DOCUMENT IS 1

UC-80

LAWRENCE IJVERMORE LABORATORY

UniversityotCalifornia Livermore.Caliiomia 94550

-NOTICE-

E.-w. no, ,y of , L , 1 1

y W..I,, T O T L V '

' " ^"imenl of

"","0)"C, k

"Winy o, rpobL f o X ' ' " " " " " " * ""

P-om. diicloml. o, " :,' P r "'""' I"*'' <'

Lw^P"%Myogg
" ' " "" """"" "<" I

UCRL-52342

EFFECTIVE MASS AND DAMPING OF SUBMERGED

STRUCTURES
R. G. Dong
MS. date: April J, 1978

WSTItlUUTIOM OF TH.IS DOCUMENT IS UNLIMITED

Contents
Abstract
j
Introduction

Structures and Excitations of Concern

Hydrodynamic Theories
Some Methods Used for Current Design Analysis
Single Isolated Members
5.1 Procedure Recommended by Newmark and Rosenblueth
5.2 Added Mass for Single Isolated Members
5.3 Effect of Finite Length on Added Mass for Single Isolated Members
5.4 Effect of Partial Submersion on Added Mass for Single Isolated Members
5.5 Added Damping for Single Isolated Members
5.6 Effect of Siiuciuial Siic oil Added Damping for Single hclatcd Member;
. 5.7 Range of Applicability of the Added Mass and Damping Concept
for Single Isolated Members
6.
Multiple Members
6.1 Complexities Associated With Multiple Members
6.2 Hydrodynamic Coupling for Groups of Cylinders
6.3 Hydrodynamic Coupling for Rigid Members Surrounded by a
Rigid Circular Cylinder
6.4 Hydrodynamic Coupling for Coaxial Cylinders
6.5 Damping for Multiple Members
7.
Conservative Choice for Added Mass
8.
Computer Codes in Current Use
9.
Conclusions and Recommendations
9.1 Idealized Single Isolated and Multiple Members
9.2 Spent Fuel Storage Racks
9.3 Main Steam-Relief Valve Line
9.4 Internals of the Reactor Vessel
9.5 Methods Used for Current Design Analysis
10. Acknowledgments
11. References

1.
2.
3.
4.
5.

,,,,,,.

I
1
3
3
8
10
10
10
19
19
19
.is
29
32
32
32

'

.,

40
47
47
62
64
64
64
65
66
66
67
71
71

Hi

EFFECTIVE MASS AND DAMPING OF SUBMERGED

STRUCTURES
ABSTRACT
Various structures important for safety in nuclear power plants must remain functioning
in the event of an earthquake or other dynamic phenomenon. Some of these important
structures, such as spent-fuel storage racks, main pressure-relief valve lines, and internal
structures in the reactor vessel, are submerged in water. Dynamic analysis must include the
force 2nd damping pff?Ct*? f water. This report provides a technical basis for evaluating the
wide variety of modeling assumptions currently used in design analysis. Current design
analysis techniques and information in the literature form the basis of our conclusions and
recommendations. We surveyed 32 industrial firms and reviewed 49 technical references. We
compare various theories with published experimental results wherever possible. Our
findings generally pertain to idealized structures, such as single isolated members, arrays of
members, and coaxial cylinders. We relate these findings to the actual reactor structures
through observations and recommendations. Whenever possible we recommend a definite
way to evaluate the effect of hydrodynamic forces on these structures.

1. INTRODUCTION
To ensure that various structures important to the
safety of nuclear power plants remain functioning
during a severe earthquake or other dynamic
phenomenon, detailed dynamic analyses must be
performed. A number of structures, such as spentfuel storage racks, main pressure-relief valve lines,
and internals of the reactor vessel, arc submerged in
water. For these structures, the effect of the water in
terms of forces and damping must be considered. A
wide variety of modeling assumptions are being used
in design analysis, and, at present, there are no
uniform positions by which to judge the adequacy of
the assumptions. The objective of this project is to
provide a technical basis for evaluating the assumptions, and to recommend suitable methods to
account for the effect of the water.
The methods investigated include the added mass
and added damping concept, current design methods, and methods underdevelopment. Experimental
results available in the literature form the basis of our
evaluation whenever possible. Following a procedure agreed upon at the start of the project, we
focus on two groups of idealized structures: single
isolated members and multiple members. The second
group includes two parallel cylinders, members near
a boundary, an array of members, and coaxial
cylinders. We relate our findings to spent-fuel

storage racks, main pressure-relief valve lines, and

the internals of the reactor vessel through observations and recommendations. Development of new
methods and performing rigorous analyses were not
major endeavors for the project.
An extensive survey of the literature and industrial
firms was carried out. Forty-nine references (1-49),
listed in the order reviewed, covered single isolated
members and multiple members. Thirty-two industrial firms were contacted (see Table 1); this survey
revealed that the design methods in current use are
quite varied and that, in some instances, rather
sophisticated developments are taking place.
Of special interest to the Nuclear Regulatory
Committee (NRC) is a recommendation for added
mass and damping values made by Newmark and
Rosenblueth.' This recommendation forms the
baseline for NRC's current position on the subject.
We compared this recommendation with the
theoretical and experimental results we reviewed.
The fluid-structure interaction for multiple members is significantly more complex than for single
isolated members and is less well understood.
Consequently, we find it advantageous to separate
single isolated members from multiple members in
our presentation!
1

Table 1. Industrial firms surveyed.

Argonnc National Lab.
Argonne, Illinois
James Kennedy
Yao Chang
S. S. Chen
Babcock & Wilcox
Lynchburg, Virginia
Arthur F. J. Eckert
Beclilel
San Francisco, California
Sidney Ting

Gilbert '
Reading, Pennsylvania
Donald Croneberger
Los Alamos Scientific Lafcorator
Los Alamos, New Mexico
Tony Hirt
Lockheed Corp.
Sunnyvale, California
Robert L. WaiJ
NASA
llehu Struck
Denny Kross

Civil Engineering Lab.

Port Huente, California
William Armstrong
Francis Liu
Dallas Meggitt

Naval Post Graduate School

Monterey, California
Targut Sarpkaya

Combustion Engineering
Windsor, Connecticut
Bob Longo

Naval Research Laboratory

Washington, D.C.
Owen M. Griffin

EDAC
Irvine, California
Robert P. Kennedy

Nuclear Energy Services Incorp.

Danbury, Connecticut
Igbal Husain

EDS
San Francisco, California
Majaraj Kaul

Nuclear Services Corp.

Campbell, California
Henry Thailer

EPRI
Palo Alto, California
Conway Chan

NUS Corporation
Boston, Massachusetts
Howard Eckert

Exxon Nuclear Co., Inc.

Richland, Washington
Denny Condotla
Charles A. Brown

Offshore Power Systems

Jacksonville, Florida
Richard Orr
Jeff Shulman

Frederick R. Harris, Inc.

New Vork, New York
Herman Bomze

Oregon State University

Corvallis, Oregon
Tokuo Yamamoto

General Electric
San Jose, California
Lun-King Liu
Bob Buckles

Physics International
San Leandro, California
Dennis Orphal

Program & Remote Systems

St. Paul, Minnisota
Donald F. Melton
Carl Ncwmcyer
Sargent & l.uncly
Chicago, Illinois
Suren Singh
Norman Webber
Stone & Webster
Boston, Massachusetts
George Bushell
Southwest Research Institute
San Antonio, Texas
Frank Dodge
Universal Analyties
Los Angeles, California
Dave Herting
University of California
Berkeley, California
Anil K. Chopra
Ray Clough
John Wehausen
URS/John Blume
San Francisco, California
Roger Skjt'i
Rrger Scholl
Wachter
Pittsburg, Pennsylvania
Mr. Wachter
Dave Secrist
Westinghouse
Pensacola, Florida
John Gormley
Tom C. Allen
George J. Bohm

2. STRUCTURES AND EXCITATIONS OF CONCERN

The nuclear power-plant structures and excitations of concern arc shown in Table 2. Seismic loads
are considered described by the response spectrum in
the NRC's Regulatory Guide 1.60 (R.G. 1.60).* The
horizontal spectrum is shown in Fig. 1. The time
histories for the pressure-relief and blowdown loads
were provided by NRC, and are shown in Figs. 2 and
3, respectively. The predominant frequencies are
indicated. The structures of concern vary in
dimension and arrangement depending on the design
and the plant. A representative list of dimensions
and/or natural trequencies are given in Tabie 3.

Table 2. Structures and excitations of concern.

Structures

Excitations

Spent-fuel storage racks

Seismic

Main steam-relief valve line

Pressure relief
Blowdown-induced loads
Seismic

Internals of reactor vessel

Blowdown-induced loads
Seismic

Table 3. Representative sizes and natural frequencies of structures of concern.

Size

Structure
Fuel elements
Fuel bundles, BWR
fuel bundles, PWR
Fuel
Fuel
Fuel
Fuel
Fuel
Fuel

racks: firm
racks: firm
racks: firm
racks: firm
racks: firm
racks: firm

' 0.5 in. D

' 5.5 x 5.5 in.
' 1 0 x 1 0 in.

(1)
(2)
(3)
(4)
(4)
(5)

Main steam-relief
valve line c

Natural
frequency

Condition

~ 3 Hz"
3 Hz"

In w-j(er
In water

17 to 33 Hz
10 to 20 Hz
6 to 9 Hz
~!2Hz
~ 10.5 Hz
~ 1.15 Hz
8 in. D, 72 in. L
8 in. D, 72 in. L
8 in. D, 396 in. L
8 in. D, 396 in. I
12 in. D, 72 in. L
12 in. D, 72 in. L
12 in. D, 396 in. L
12 in. D, 396 in. L

Reactor core barrel

0.5
1.2
0.02
0.04
0.8
1.8
0.03
0.06

H/
Hz
Hz
Hz
Hz
Hz
Hz
Hz

40 Hz*
10 Hz"

Full and in water

Full and in water
Full and in water
Full in air
Full in water
Full in water
In air
In air
in air
In air
In air
In air
In sir
In air
In air
In water

"Approximate frequency values provided by NRC.

The industrial firms generally wished to remain anonymous.
c
Range of diameters and lengths provided by NRC.

3. HYDRODYNAMIC THEORIES
Five variations of hydrodynamic theory seen in the
literature are listed in Table 4, approximately in the
order of increasing complexity. For the dynamic
effect on submerged structures, the two simplest
theories are used most. The incompressible invicid

theory, sometimes referred to as potential theory, is

used for nonflexible members; i.e., members that can
be treated as translatingrigidbodies. The compressible invicid theory is used for flexible members, such
as flexible coaxial cylinders.

'OJ>

1(T
Frequency cps

Main steam-relief
valve lines

Reactor core barrel

L-Spent-fuel storage racks
L

Fig. 1.

- Fuel bundles

Horizontal design response spectra, scaled to I-g horizontal ground acceleration (from R. G. 1.60).'

Bubble pressure
Theoretical

Period range = 0.115 to 0.144 s

Frequency range = 8.7 to 6.9 Hz
<>
Pressure in water 10 ft
from bubble center

Theoretical

0.134-

0.144

0.2

0.115-

- 4

'.
Time s

Fig. 2.

Pressure in bubble and water, steam relief.

0.6

Axial elevation: 2
Force direction: X
| 0.08 - ~ | 0 . 0 7 9 1 * - 0 . 0 7 6 -{0.072 - ~ j 0.075 -~|

Average period = 0.076 s

Average frequency = 13.2 Hz
-1.2
0.1

0.2

0.3
Time s

Fig. 3.

Blowdown excitation; horizontal force on core barrel.

0.4

0.5

liable 4.
Theories

Hydrodynamic theories.

Applicable conditions

Incompressible
invicid

Virtually no boundary layer

--C
(Potential theory)

'

Fluid escape is easy

/////////////////
O

Slci

V//V/V//V///V//A

Compressible invicid

Virtually no boundary layer

Fluid escape is not easy

Incompressible viscous

Appreciable boundary layer

Fast
Fluid escape is easy

77777777777777777"

Compreisible viscous
(Navier-Stokes)

Appreciable boundary layer

77777777777777777Fluid escape is not easy or

velocity is high

0-

Fast

rpv//////////////

Nonlinear

Appreciable boundary layer

Velocity is very high

Very
fast

4. SOME METHODS USED FOR CURRENT DESIGN ANALYSIS

Our survey of industrial firms revealed a variety of
methods (see. Table 5), for calculating added mass
and damping. The firms' identifications are kept in
confidence, as was desired by a majority of those
providing information. Some overlap exists, so that
each method shown may represent more than one
firm. The philosophy behind each method is
illustrated using a simplified representation of a fuel
bundle with its can enclosure. For clarity, we show
only four fuel elements per bundle, although in
reality a typical fuel bundle has from 60 to 200 fuel
ek--mentj>. A single fad bundle with its can is an
example of a single isolated member in lable 5, and
two or more fuel bundles are examples of multiple
members. The volumes of water included in the
calculations of virtual mass are shaded in Crosshatch.
In the case of fuel bundles, the mass of the water
within the can is simply taken as part of the structural
mass. The mass of a certain volume of water outside
of the can is added to the structural mass, and this is
commonly referred to as the added mass from
submersion. The methods used for calculating this
added mass is quite varied, as indicated in Table 5,
and they are largely based on engineering judgment
together with whateve. analytical and/ or experimental information was available at the time.
A detailed description of the basis for each method
was not provided by the firms contacted; perhaps for
most, the only basis was engineering judgement. In a
few cases, references were cited; however, we found

Table 5.
Method
number
1
2

no direct relation between the methods and the

references.
Table 5 is self-explanatory for most of the cases
shown. In method 5, the procedure presented by
Fritz7 for coaxial cylinders was used to approximate
the interaction between the central member and the
eight peripheral members of a 3 x 3 array. In method
9, the cans are in contact with each other, so that
virtually no water exists between adjacent cans.
The bases for the damping value used are likewise
quite varied. Zero damping was chosen in some cases
to ensure conservatism. In some instances, hr
structural plus added damping was taken as 2 to 2-'/2
times the structural damping. The basis for this
appears to be various references, such as 3 and 17,
which choose to present experimental results for
total damping in terms of a factor, such as 2, times
the structural damping. We disagree with this interpretation, for it implies that the submerging water
somehow knows how much damping is in the structure, and it subsequently adds an equal amount. The
experimental results given in Ref. 3 and 17 could
just as well be expressed in terms of an added
damping, which we feel is a more valid interpretation. It is our opinion thai the use of a factor times
the structural damping, as used in method 3 and
considered for use in methods 7 and 8, should be
discouraged.
Our assessment of the methods described in Table
5 is given in Section 9.5 of this report.

Design metho ds for evaluating added mass in current use for seismic excitations.

Single, isolated member

Potential theory
(perfect fluid)
Potential theory

Multiple members

Damping

Potential theory
Potential theory
modified by experiments

ID

Enri
12
y/

.'A

YA

2 times structural
damping

Tabic 5. (continued)
Method

number

Single, isolated member

Multiple member

oo
oo

Where

Dumping

o
o

""]

Added damping = 0%

40'/; of

Potential theory

Added damping
= 0 to 3%

l-'rit/. (Kef. 7)

lor natural fri-i|uency evaluations

Y7J7?

^H2

Added mass
= displaced water

Added damping = 2%

l o r incrtial load evaluations

Use the smaller of:

Actual measured amount of water

surrounding the racks

Evaluate the added mass as if the cans

were single and isolated

Prefer to use 2 to
2-'/j x structural damping, but
are using added damping of 2%
per NRC's request

Added mass
= displaced water

Added damping = I sually ()%,

but might consider using 2 times
structural damping

Added damping = 0%

Very thin film of water

An assessment of validity of these methods is given in Section 9.5 of this report. A method recommended by LLL, not shown in (his table,
is explained in Sections 9.1 and 9.2 of this report.

5. SINGLE ISOLATED MEMBERS

5.1

Procedure Recommended By
Newmark and Rosenblueth

A procedure for evaluating the added mass and

damping of single isolated members submerged in a
fluid was suggested by Newmark and Rosenblueth.5
For added mass, they suggested:
"if the structure is a ionji, rigiii piisui uit flexible

supports, moving in a direction perpendicular to its

axis, flow of liquid around the structure isessentially
two-dimensional. Under these conditions, the added
mass is that of a circular cylinder of liquid having the
same length as the prism and a diameter equal to the
width of the projection of the prism on a plane
perpendicular to the direction of motion (Fig. 4)."
For added damping they said:
"... damping due to liquid viscosity may be
disregarded. Energy dissipation due to radiation into
the liquid may be more important, but the model
tests to which we have referred [ 1 ] indicate that it will
not exceed about 2% of critical for submerged
structures of ordinary dimensions."
An evaluation was carried out and included in our
presentation.

5.2

Added Mass for Single

Isolated Members

If a single isolated member is accelerated in a

stationary fluid, its acceleration induces the fluid in
its immediate neighborhood to accelerate. The
accelerating fluid in return induces an added mass
effect onto the member. Under sufficiently small
amplitudes of motion, cyclic or unidirectional, the

Fig. 4. Submerged body and its virtual mass/

10

added mass phenomenon can be described in terms

of an added mass coefficient Cm defined as
added mass of fluid
C

=
reference fluid mass '

where the reference fluid mass is that ol the cylinder

of fluid of diameter equal to the dimension perpendicular to the direction of motion, or, in some cases,
ii is the mass of the displaced fluid. The added mass
phenomenon for single isolated members has been
rather extensively investigated experimentally and
analytically. Theoretical treatment has been quite
successful using the potential theory.
Experimental data for single isolated members are
available in Refs. 1, 3, 4, 6, 12, 13, 17, 20, and 29.
Potential theory results are given in Refs. 2 and 7,
Table 6. The available experimental data, potential
theory results, and Newmark and Rosenblueth's
(N&R) recommendations are compared in Figs. 5a
through 5i for a variety of specimen geometries.
N&R's recommendation as worded in Ref. 5 applies
only to the situations of Figs. 5a, 5c, and 5d; therefore, comparison with the recommendation is carried
out only for these three cases. Notice that the added
mass coefficient is independent of the cross-sectional
geometry of the specimen for N&R's recommendation. This is a simplification embodied in the recommendation, which is important to keep in mind, for
we will see later that it will give rise to some uncertainties about conservatism when using the recommendation.
A comparison between Figs. 5a and 5b and
between Figs. 5e and 5f reveals that the value of Cm
for a fluid moving around a stationary specimen is
higher than the value for a specimen moving in a
stationary fluid. This higher value of Cm is exhibited
theoretically,6" as well as experimentally, and is
important to account for in real applications. For a
stationary circular cylinder in a moving fluid, Cm =
2, which means the hydrodynamic force acting on the
stationary cylinder is twice the mass of fluid
displaced times the acceleration of the fluid. By
comparison, for a translating circular cylinder in a
stationary fluid, Cm = 1, which means the total force
required to accelerate the cylinder is the mass of the
cylinder plus the mass of the displaced fluid
multiplied by acceleration of the cylinder. For the
case in which both the cylinder and fluid are in
motion, these two force contributions should be calculated separately and superimposed.

Table 6.

Two-dimensional bodies.

Section through body

Translational
direction

Hydrodynamic mass
per unit length

Vertical
m h = 1 7T,9a2

Vertical

m n = 1 Tip a

Vertical

\l

I II

1 Tip a.

Vertical

I, i\

mj, = 17rpa
2a
Vertical
a/b =

ni], = 1 n p a

a/b = 10

mh = 1.14 up a 2

a/b = 5

m h - 1.21 77pa 2

a/b = 2

mj, = 1.36 7:pa 2

a/b = 1

m h = 1.51 Tip a 2

a/b = 1/2

m h = 1.70 Tip a 2

a/b = 1/5

m h = 1.98 rcpa2

a/b = 1/10

mj, = 2.23 Tl p a
Vertical

v^
^
-2a*

d/a = 0.05

m n = 1.61 ffpa

2a

d/a = 0.10

m h = 1.72 Tip a 2

d/a = 0.25

mj, = 2.19 Tip a 2

1
^S

Table 6. (continued)
Translational
direction

Section through body

Hydrodynamic mass
per unit length

Vertical
a/b = 2

mh = 0.85 7Tpa2

a/b = 1

mj, = 0.76 7T pa

a/b = 1/2

mh = 0.67 7Tpa

a/l> - 1/5

mn - 0.61 pa

Vertical
(normal to free
surface)

mh = 0.75 it p a

a/b = 1

Horizontal

(parallel to free
surface)
ni), = 0.25 npa

a/b = 1

a/b = 1;
e/b =

Vertical
(normal to free
surface)

mn = 0.75 7rpa:

e/b = 2.6

mn = 0.83 7rpa'

e/b = 1.8

mn = 0.89 7rpa:

e/b = 1.5

mh = LOOffpa

e/b = 0.5

mn = 1.35 7rpa'

e/b = 0.25

mh = 2.00 7T pa

\\\\\\\\\\\SXSN\
Vertical

J . _

-T
b

V- 2a~| T

2b
I

a/c = 2.6
a

/ c = 2.6
b/c = 3.6

b/c = 3.6

mh = 2.11 ffpa

Table 6. (continued)
Three-dimensional bodies
Body shape
1. FLAT PLATES
Circular disk

Translations!
direction

Hydrodynamic mass

Vertical
mh = 8- pa3
Effect of Frequency of
Oscillation on
Hydrodynaniic Mass of
a Circular Disc
CO = angular frequency
c = velocity of sound
in medium

Nondimensionul frequency - u

Elliptical disk

As shown

mh = Kba2 - p
n
6
b/a K
<*> 1.00
14.3
12.75
10.43
9.57
8.19
7.00
6.00
5.02
4.00
3.00
2.00
1.50
1.00

Rectangular plates

0.991
0.987
0.985
0.983
0.978
0.972
0.964
0.952
0.933
0.900
0.826
0.748
0.637

Vertical
mh = K V p b
4
b/a
1.0
1.5
2.0
2.5
3.0
3.5
4.0

0.478
0.680
0.840
0.953
1.00
1.00
1.00
1.00

Table 6.
Body shape

(continued)
Hydrodynamic mass

Translatlonal
direction

Triangular plates

Vertical

9 3 (TAN0j? fl
r m = a (77)

2. BODIES OF REVOLUTION
Spheres

Vertical

2a

Ellipsoids

Tnpa

Vertical
mh = K '

a/b
1.00
1.50
2.00
2.S1
2.99
3.99
4.99
6.01
6.97
8.01
9.02
9.97

Kfor
axial
motion
0.500
0.305
0.209
0.156
0.122
0.082
0.059
0.045
0.036
0.029
0.024
0.021
0

u2

rrpab
Kfor
lateral
motion
0.500
0.621
0.702
0.763
0.803
0.860
0.895
0.918
0.933
0.945
0.954
0.960
1.000

Table 6.

(continued)

Body shape

Hydroclynamic mass

Translations!
direction

Approximate method for elongated bodies of revolution.

K,pV = Ke fl + 17.0 (cp - j )* f 2.49 (M - - | - ) ' + .283 [('<) ~ j)'

where;

('1 " j)']

Kj Hydrodynamic mass coefficient for axial motion

Kc - Hydrodynamic mass coefficient for axial motion
of an ellipsoid of the same ratio of a/b
V Volume of body
C n - Prismatic coefficient

4V
b 2 (2a)

M - Nondimensional abscissa
to maximum ordinate

"m

corresponding

'

IQ, r j - Dimensionless radii of curvature at nose and tail

Ro (2a)
r

0 =

Rl (2a)
r

l =

Lateral
motion

Munk has shown that

the hydrodynamic mass
of an elongated body
of revolution csn be
reasonably approximated
by the product of the
density of the fluid, the
volume of the body, and
the k factor for an
ellipsoid of the same
a/b ratio.

Table 6.
Body shape
Sphere near a free surface

(continued)
Translational
direction
Vertical

Hydio-.lynamic mass
2
3
mh = K 7T/aJ
g/2a
0
0.5
1.0
1.5
2.0
23
3.0
3.5
4.0
4.5

Ellipsoid near a free surface

0.50
0.88
1.08
1.16
1.18
1.18
1.16
1.12
1.01
1.00

Vertical
4

mh
a/b

,.2

K 7T/aab2
3
2.00

s/2b
1.00
2.00

0.S13
0.905

3. BODIES OF ARBITRARY SHAPE

Ellipsoid with attached
rectangular flat plates

Vertical

4
.1
mj, = K tfpab
a/b = 2.00; c = b

.'

J'

c.d= N 7T ab
N

0
0.20
0.30
0.40
0.50
Ellipsoid with attached
rectangular flat plates
near a free surface

Vertical

0.7024
0.8150
1.0240
1.1500
1.2370

mh =

7rpab2

a/b = 2.00; c = b
c.d = N 7T ab
N

0
0.20
0.30
0.40
0.50

0.9130
1.0354
1.3010
1.4610
1.5706

'

Table 6.
Body shape

(continued)
Translations!
direction

Streamlined body

Vertical

Hydrodynamic niass

mh = 1.124 p

d =

= 2.38
b

c +b

= 2.11
c

Area of horizontal "tail" = 25% of area of body maximum horizontal section.

Streamlined body

Vertical

'^J-

mh

0.672 pT 7rad2

2b

1
c +b

2a

= 3 .4

Area of horizontal "tail" - 20% of area of body maximum horizontal section.

"Torpedo" type body

Vertical
m h = 0.818 irp\r(2a)

\%z

31

2b

-"2a-

-S-s.0
b

Area of horizontal "tail" = 10% of area of body maximum'horizontal section.

Table 6. (continued)
Body shape

Iranslalional
direction

V-Fin type body

Hydrodynamic m.iss

Vertical
mh = .3975 p V

4=1.0
h

Parallelepipeds

= 2.0
c

Vertical
mh = K p a ' b

J8

b/a

J
2
3
4
5
6
7
10

2.32
0.86
0.62
0.47
0.37
0.29
0.22
0.10

Some scatter is seen in the experimental data.

Various possible effects contributing to the scatter
includes specimen flexibility, frequency dependency,
amplitude dependency, and normal experimental
variability. Specimen flexibility appears to be an
important factor that tends to lower the added
mass." The cantilevered and the thin-walled
specimens exhibited lower added masses than do
spring-mounted rigid specimens. Taking into consideration this lowering effect from specimen
flexibility, the agreement between experiments and
potential theory can be considered quite good. The
exception is Fig. 5i, where we suspect the theoretical
resuits reported in Rcfa. 2 aiid 7 arc incorrect. Our
reasoning is that Cm for a cube and a sphere should
be similar. Yet, Fig. 5i gives Cm = 2.32 for a cube
(b/a = 1), while Fig. 5e gives Gm = 0.5 for a sphere.
The experimental value for a cube shown in Fig. 5i is
Cm = 0.67. This compares much more favorably with
Cm for a sphere than with the theoretical value for a
cube, which supports our contention that the
theoretical results of Fig. 5i are incorrect. To our
knowledge, the case of Fig. 5i is the only error
contained in Refs. 2 and 7; however, some caution
might be exercised in using these references for cases
in which no experimental data is available ifor
comparison.
Trie good agreement between potential theory and
experimental data lead us to conclude that potential
theory satisfactorily describes the added mass
phenomenon. This confirms the opinion, as expressed in Refs. 15, and 16, that for single isolated
members the compressibility and viscous effects of
the water are negligible compared with inertial
effects. Potential theory, while unable to model
compressibility and viscosity, can model inertial
effects quite well. Taking the position that the
added mass given by potential theory is valid, a basis
for evaluating the adequacy of N&R's recommendation becomes possible. In the case of a circular
cylinder moving in a stationary fluid, N&R's recommendation coincides with potential theory, Fig. 5a.
In the case of a rectangular cylinder moving in a
stationary fluid, N&R's C m value can be less than
(Fig. 5c) or greater than (Fig. 5d) that given by
potential theory, depending on the direction of
motion of the specimen as illustrated by Figs. 5c and
5d. The difference between N&R's Cm values and
those given by potential theory can be quite
significant. For example, consider a cylinder of
square geometry; i.e., a/b = 1 . 0 . In Fig. 5e, the
theoretical Cm value is 1.5 compared with N&R's
C m value of again 1.0. We will show later that to help
assure conservatism we would want to maximize Cm
under some conditions and to minimize Cm under

others. Because using N&R's recommendation can

result in Cm values either greater than or less than
the theoretical Cm values, conservatism is not
necessarily ensured by following this recommendation. The use of potential theory to evaluate added
mass is preferred in terms of greater control over the
conservatism as well as providing greater general
accuracy.

5.3 Effect of Finite Length on Added

Mass for Single Isolated Members
/'
In the case of a finite length member, the fluid
flows around the end(s) as well as around the length.
Therefore, the inertial resistance to motion is less
than that for an infinitely long member. Figures 6
and 7, from Ref. 4, illustrate the effect experimentally and theoretically for specimens with both
ends free for fluid to flow around. These curves could
apply in an approximate sense to cross sections other
than those of the figures.

5.4 Effect of Partial Submersion on

Added Mass for Single Isolated
Members
The added mass effect decreases near the water
surface for a partially submerged member. The
added mass distribution based on potential theory is
shown in Fig. 8 for a vertical circular pier for three
levels of partial submersion.5-' The decrease in total
added mass as a function of depth of submersion is
given in terms of a correction factor in Fig. 9. Experimental results for vertical cylinders are shown in Fig.
10 with the partial submersion given as a fraction of
the total length.3 Unfortunately, the specimen length
was not given, so that no comparison can be made
with Fig. 9. The general trend, however, agrees with
that of Fig. 9. The correction factors of Fig. 9 could
apply in an approximate sense to vertical cylinders of
cross sections other than circular.

5.5 Added Damping for Single Isolated

Members
The damping force acting on a submerged member
is usually relatively small and not included in analysis
as an acting force. Instead, the effect is usually
19

_ Added mass of water

Reference water mass

Ref. water mass = Cylinder of water of diameter

equal to dimension perpendicular
Lo u i i e c i i u i i 01 i i i u i i o n UIII'CJ-

otherwise noted

Curves'.
Potential theory (Refs. 2 and 7)
Newmark and Rosenblueth's recommendation (Ref. 5)
Experimental results:

Fig. 5.

O
D

Solid on springs
Cantilevered beams

Cantiievered beams

<

Solid on springs

>

Fixed solid in flow

Fixed solid in oscillating fluid

Solid oscillating in fluid

Thin-walled hollow beams

Solid oscillating in fluid

Fixed solid in oscillating fluid

~Orr
#

fv,

fy,

Ref. 1
Ref. 1
Ref. 3

*$&i

Ref. 4

-<> -*Q"~ -*-~.jL

JV.

Ref. 13
Ref. 6
Ref. 12
Ref. 17

""
_. J*r7". i_

Ref. 20
" e ^- 29

Comparisons of the potential theory, Newmark and Rosenblueth's recommendations, and experimental data.

1.5

"V

Potential theory
/

1.0

o
0.5

Thin- '
tf walled
. tube

17

(a)

N & R's recommendation

20

Refs.

Fig. 5a. Comparisons for'un oscillation circular c} tinder in still fluid.

3.0

(b)

2.0

C,m
1.0 -

13

29

Refs.

Fie- Sb. Comparisons for a fixed circular cylinder in oscillating fluid.

Fig. 5c.

Comparisons for an oscillating rectangular cylinder in still fluid.

1,5r

(d)

1.0

m
0.5

3
a/b

Fig. 5d.

Comparisons for an oscillating rhombic cylinder in still fluid.

21

(e)

1.0 i-

48 &

C m 0.5

Added mass
Mass of displaced Q
fluid

12

Fig. Se. Comparisons for an oscillating sphere in still fluid.

(f)
2.0
1.5
m 1.0

0.5
Added mass
o
Mass of displaced
fluid

29
Ref.

Fig. 5f.

t
J

Comparisons for a fixed sphere in an oscillating fluid.

(9)

2.0
C

m 1.0
6
Ref.
Fig. 5g. Comparisons for a fixed infinitely long plate in oscillating fluid.

1bI
10
A

cm =

a2
a

7rP -b
4

^r

"

(h)
A

m 0.5

Added mass

I
5

b/a

Fig. 5h.

Comparisons for an oscillating, finite plate in still fluid.

(i)

3.5 ni

"I

from
\ >^Calculations
Ref. 2 in error?

\V

C_
m =

Added mass 1 -0 Mass of displaced

fluid
0.5

.....|

5
b/a

Fig. Si.

Comparisons for an oscillating solid in still fluid.

22

O Rectangular plate
Cylinder
X Theory
-

8 4
2
1
Ratio of length to diameter or width
Fig. 6. Circular cylinder! and rectangular plates/

0.01
0.1
Relative length

1
10
thickness to width

F=

T77777r^77777777M777777
Fig. 7. Relative effec* of virtual mass in parallelepipeds
square side moving broadside on.4

described as an equivalent viscous damping. The

contributions to added damping are:
Fluid viscosity.
Component impact.
Wave generation.
Acoustic generation.
The last two are forms of radiation damping; i.e.,
wave or acoustic energy generated radiate away from
the submerged member. We do not expect a
significant amount of acoustic energy generation for
the structures and excitations of concern. Wave
generation is generally not important for fully
submerged structures under seismic excitation,1-5
and it seems reasonable to extend this to other types
of excitations, such as vibrations induced in the main

- = ir

vA
Fig. 8. Added mass distribution for a partially submerged
member.8'*

steam-relief valve line by normal pressure relief. For

partially or fully submerged members in a finite-size
water enclosure, radiation damping is again, usually
not taken into consideration, because the radiation
energy may bounce off the enclosure walls back to
the submerged member. Therefore, in the actual
structures of concern, we choose to ignore wave
generation as a source of damping. Component
impact may be a significant source of damping for
23

5 1.0
4 0 . 9 cm
(1)

Circular section
-1.7 cm-

0.3 cm

Fig. 9.

multiple members, and this will be discussed further

when we address multiple members later in this
report; however, it is not a source of damping for
single isolated members. Therefore, for single
isolated members, fluid viscosity is the only source of
damping in need of consideration.
Added damping is not as thoroughly investigated
in the literature as was added mass. Theoretical
predictions of damping are seldom attempted
because experimental values are usually more
reliable. Therefore, our conclusions on added
damping are based whenever possible on published
experimental data.
Experimental data on added damping are represented in Refs. 1, 3, 20, and 27 covering a variety of
specimen shapes and experimental conditions. The
most extensive set of data is found in Ref. 20 where
circular cylinders of 0.31-, 0.5-, 0.75-, and 1.0-in.
diameters are investigated over the frequencies 2.5 to
18.6 Hz and amplitude-to-diameter (A/D) ratios of
up to 2.0. The data of Ref. 20 indicate that viscous
damping applies up to an A/ D value of 0.32 for the
smallest specimen (0.31 in. diam) and 0.5 for the
largest (1.0 in. diam). Beyond the viscous damping
range is the nonlinear range where the damping force
becomes proportional to the square of the velocity.
The change from linear to nonlinear behavior with
increasing A/ D value was quite distinct as indicated
in Fig. 11 for two examples from Ref. 20. Nonlinear
damping is seen to be greater than linear damping, so
that using the linear damping value as an approximation in the nonlinear range will be conservative.
The A/D values where the change from linear to
nonlinear damping occurs are plotted vs. specimen
diameter in Fig. 12. A gradual increase with diameter
is seen; however, without data for larger diameters
we are uncertain if the trend would continue to
increase for the sizes of actual structures of concern.
In structural analysis, damping is expressed as
either .a damping coefficient or a percent of critical

IN.

Liquid mass correction factor in circular pier.'-'

(2)

24

E
u

(3)

Square section

Rectangular section
c
o

5 cm
(4)

Pier section

Cross section of various models

o

CD
C7
L-

(D

E
at
H-

SI

+->

|"0.5h

1
2
Virtual mass coefficient

Fig. 10. Variation of virtual mass vs depth of submergence.1

damping. Damping coefficient describes the damping independent of the mass and stiffness of a
structure, whereas, the percent of critical damping is
a description associated with the mass and stiffness.
To see which description best fits the added damping
from water we converted the data in Ref. 20 to both
an added coefficient and an added percent of critical

0.5
Diameter in.
Fig. 12. Air.plltudc/dlamcter value for linear damping, oscillating submerged circular cylinders.10

a
Fig. 11. The calculated viscous damping coefficient C, ( or, /})
vs the dimensionless amplitude a = A / D . The calculated points
are denoted by the "+" symbol and the inherent error bounds on
these points are inclosed by the "( )" symbol. The solid
curve is the two-segment straight line fit to the calculated points.30

damping. These are plotted vs! frequency in Figs. 13a

and 13b, respectively. The added coefficient varied
significantly with frequency in an inconsistent
manner; i.e., the 0.13-, 0.5-, and 0.75-in. specimen
showed an increase with frequency, whereas the 1.0in.-specimen showed a decrease. The added percent

of critical damping varied less and was more

consistent; i.e., all specimens showed cither a
constant value or decreasing value with frequency.
Those showing decreasing values are the 0.31- and
1.0-in. specimens, as indicated by dashed curves in
Fig. 13b. At this point, we chose to impose a simplification to proceed with forming a workable recommendation. Therefore, we chose to describe the
percent of critical damping as a constant with respect
to frequency. The constant values are indicated by
the solid curves in Fig. I3b, and are plotted in Fig. 14
as a function of specimen size. Experimental data
from Refs. 1,3, and 27 are added, and these included
specimens of circular, square, and plate crosssections. The data points are few, and the scatter is
moderate; yet a general trend is apparent in, that
added damping decreases with increasing specimen
size. The trend was established by the data for
circular specimens and was not contradicted by the
data for square and plate specimens. Further
discussion of this trend follows.

5.6

Effect of Structural Size on Added

Damping for Single Isolated
Members

The decrease in added damping with increasing

structural size indicated in Fig. 14 is further
confirmed by comparing with a similar trend
established for damping of water sloshing in pools.
The latter trend is well established, and expressions
for the dependence of damping on pool size are given
25

Added damping In terms of damping coefficient.

8
10
12
14
Natural frequency in air - Hz

16

Added damping in terms of the percent of critical damping.

Fig. 13. Damping data on oscillating circular cylinders in still water."

18

20

Frequency
Test
Cross sections Refs. config. Mode range
O
D <=>

O
2 to 8 Hz
20
0

-o

-a

CO

O
35

P A
^ A

O-

D-

2 to 4 Hz

15 to 25 Hz

2 to 7 Hz

16 Hz

I/I

E
g 12

27

Circular
Square
Plate

|10
3
T3

o
a>

O
a
to

8 -

Fig. M.

2
3
Specimen size, diameter or width in.

Percent of added damping for various specimen cross sections and sizes.

in Refs. 51,52, and 53. Several different expressions

are seen in these references; however, they all have a
common form of,
log (6) = log (A) - (q) log (tt) ,
where 8 is the damping of the water sloshing in the
pool, A is a constant, q is the constant defining the
dependence on the pool size, and R is the length of
the pool. Depending on the expression used, the
value of q ranged from 0.75 to 1.0.

To see if the decrease in added damping for

submerged single isolated members follows the trend
established for water sloshing in pools, the data in
Fig. 14 is replotted in Fig. 15 in terms of log
(damping) vs log (specimen size). The scatter is rather
wide; however, the decreasing trend is apparent. A
straight line was least-square fitted to the data as
shown.Theslopeof this line gave q = 0.85.Because the
value of 0.8> fell between the values of 0.75 and 1.0
established, for pools, we interpret this as good
27

Frequency
Test
Cross sections Refs. config. Mode range
0
D <=>

O
2 to 8 Hz
20
#
9
O
-A

P P A

O-

D-

-O

10

-D

?
7 h

3
27

D-

#
*

&
*

2 to 4 Hz

V?

l b to 25 Hz

2 to 7 Hz

16 Hz

A
*

P
P
A
P

A
-A
-O
-D
0.3

0.5

0.7

Specimen size in.

Fig, IS. A log vs log plot of the percent of added damping for various specimen cross sections and sizes.

10

confirmation that added damping for submerged

single isolated members decreases with increasing
structural size and that the dependence on structural
size is reasonably characterized by Fig. 15. Our
findings were shown to D. D. Kana of the Southwest
Research Institute,** and he agreed that our
treatment and interpretations are reasonable in view
of the current state of understanding of the added
damping phenomenon. We, then, used Fig. 15 to
extrapolate the added damping values for the
structural sizes of concern shown in Table 3. The
results are given in Table 7. Except for single isolated
fuel elements, the damping for all other single
isolated members of the si.iuci.uiea shown in Tabic 7
is quite iow. These low values are in agreement with
Newmark and Rosenblueth's suggestion5 (See Section 5.1 of this report) that "damping due to liquid
viscosity may be disregarded" for single isolated
members of common structural sizes.

Range of Applicability of the Added

Mass and Added Damping Concept
for Single Isolated Members

some degree beyond its applicability range. Again in

Section 5.4 of this report, the added damping values
for single isolated members of structural sizes of
actual concern are generally quite low. (See Table 7).
Except for the single isolated fuel element, we may
choose to ignore the added damping. In this case, the
range of applicability for the added damping concept
may be disregarded.
Turning now to the range of applicability of the
added mass concept for single isolated members,
experimental results for added mass coefficient Cm
over a wide range of amplitudes for various specimen
geometries are shown in Figs. 16 through 19. The
aLiciisa in all four figure:; ic U...T/D v/hich is simply
2n times the A/U rr.tio; Um is the velocity
amplitude, and T is the period of oscillation in
seconds per cycle. Comparing these curves with the
theoretical value for Cm obtained using potential
theory we defined the range of applicability as the
range of Um T/D corresponding to experimental
values of Cm within 10% of the theoretical value.
The resulting A/ D values for the applicable range of
added mass are tabulated in Table 8. The A/ D values
for the applicable range of linear damping are also
indicated.

The range of applicability of the added mass and

added damping concept for single isolated members
can be considered as being defined by the smaller of
either the range for added mass or the range for
added damping. In Section 5.4 of this report, the
range of applicability of added linear damping was
found to vary from an amplitude to diameter (A/ D)
ratio of 0.32 for a 0.31-in. diameter specimen to an
A/ D value of 0.5 for a 1.0-in. diameter specimen. The
applicable range increases with specimen size, so that
we would expect the range to be greater than an A/ D
value of 0.5 for specimen diameters Jargerthan 1.0 in.
Also in Section 5.4, we observed that beyond the
range of applicability for linear damping, the
damping increases. Therefore, using linear damping
beyond the linear range would be conservative.
Consequently, depending on the degree of conservatism desired, linear added damping may be used to

The four values of A / D in Table 8 for the

applicable range of the added mass concept are in
good agreement. The A / D value for a sphere is
expected to be higher than that for a cylinder or plate
because it is a finite length specimen (more
streamlined), and, therefore, potential flow can be
expected to apply over higher values of displacements. The A/D range of 0.32 to 0.5 for the
applicability of linear damping is not drastically
different from the 0.8 and 1.4 values for added mass.
Therefore, we consider the ranges for both to be
mutually supportive.
Because the smallest structure of concern shown in
Table 7 is 0.5-in. diameter, the range of applicability
for the added mass and added damping concept for
these structures, as single isolated members, can be
considered to be no less than an A/ D value of 0.4
according to Table 8. If added damping should be
ignored, then the range would be an A / D value of

5.7

Table 7.

Added damping values projected in Fig. IS for single, isolated structures

Structure

Size, in.

Fuel elements
BWR fuel bundle
PWR fuel bundle

~0.5D
~ 5.5X5.5;
~ 1 0 X 10
8D
12 D

Main steam-relief valve line

Added damping,
% of critical

<4.2
<0.55
<0.33
<0.40
<0.29

29

Diameters
6.35 cm
5.08
4.45

i
3.81 cm
a3

-18
A 2.54

Ref.6
,

J_
20

10

J_

j_

j_

30
U

40

50

m /

Fig. 16. Mass coefficient vs (he period parameter for a cylinder."

2.0

'

Cm
1.5 J * * > > > ,

'
10.10 cm

Diameters

'
n 6.70 cm

QQfi

o 44R

7.62

2.86

1.0 -

10

20

30

40

Urn
Fig. 17.

Mass coefficient vs the period parameter for a sphere.2'

labie 8. Applicable range of motion amplitude determined from added mass and added damping
experimental data
Type of data

Specimen geometry

A/D=UraT/2JtD

Added mass

Circular cylinder"
Sphere"
Circular cylinder*
Plate'

0.8
1.4
0.8
0.8

Added damping

Circular cylinder"
diameter:
0.31 in.
0.5 in.
0.75 in.
1.0 in.

0.32
0.4
0.43
0.5

*Um
T
D
A

30

= maximum oscillating velocity

= period in seconds/cycle
= specimen diameter or width
= maximum oscillating displacement

5.1

125
UmT/D

Fig. 18.

Variations of the mass coefficient of cylinder.6

125
Um
m T/D

Fig. 19.

Variations of the mass coefficients of plates.

31

0.8. In cither case, however, the applicable range

would probably adequately cover any response to
seismic and normal steam-relief excitations. The
response to an accident condition may or may not be
within the applicability range depending on the
structure involved and its location relative to the
accident site.
Our conclusions are based on experimental data
from rather small specimens, sizes up to 3.0 in. in
diameter, and on rather low frequencies (0.35 Hz for

Ref. 29, 0.48 Hz for Ref. 6, and from 2.5 to 18.6 Hz

for Ref. 20). Theoretical considerations indicated in
the added mass coefficients should be size and
frequency independent. We made a simplifying
assumption that added damping is frequency
independent, and we developed a technique to
describe added damping as a function of structural
size. Although we feel comfortable with our
assumptions and developments, some additional
experimental verification at higher frequencies and
with larger specimens would be highly desirable.

6. MULTIPLE MEMBERS
6.1 Complexities Associated with
Multiple Members
The fluid dynamic effects on multiple members are
more complex than for a single isolated member. The
arrangement of the members, space between members, motion of one member relative to another, and
the generation of lift forces are all additional
important considerations. Added mass forces are no
longer necessarily in line with the direction of
motion, and lift forces may be generated which tend
to act perpendicularly to the direction of motion, WMM
Damping tends to be higher than for single isolated
members, and tight spaces between members, in
particular, can increase the damping measurably. 2'.47.4i Multiple-member response, in general, is
not too well understood. Current interest appears
high as evidenced in recent publications, particularly
relating to nuclear reactors. Many highly theoretical
works are presented; some are rather complicated in
terms of practical, everyday use in design analyses.
Some experimental data are available to validate
certain, often limited, aspects of the theoretical
solutions. In general, additional experimental validation is needed, and the range of applicability of
the various analytical techniques needs to be
established.
Although many of the investigations are motivated by reactor internal concerns, the results
published so far apply at best only to normal reactor
operations and not to conditions associated with a
blowdown accident. The flow rates and/ or component motions are assumed small. Conditions
associated with a blowdown accident are very likely
beyond the range of applicability of the various
techniques presented.
3 2

".

. """"

'

For our presentation, we separate our findings

with respect to three types of structural arrangements:
(1) groups of cylinder, such as arrays,
(2) groups of cylinders, or a single cylinder, surrounded by a large circular cylinder, and
(3) coaxial flexible cylinders.
The first category can apply to the main steamrelief valve line next to the pressure suppression pool
wall, an array of fuel elements in a fuel bundle, an
array of fuel bundles in a spent-fuel storage rack, and
an array of storage racks in a spent-fuel storage pool.
The second and third categories can apply to the
reactor-vessel internals.

oi

c^ dcts

6.- Hydrodynamic Coupling for Groups

Closed form solutions using potential theory are
presented in Refs. 10,18,19,32,33,34,35,36,42,43,
and 48 for added mass and lift forces for a group of
cylinders. The solutions are given in terms of multiple
summations and infinite series. The analyses are
rather complicated but quite general; a group of
different sized cylinders arranged arbitrarily can be
handled, at least theoretically. A clear physical
interpretation of the complex solutions is not
immediately apparent. Some insight is provided in
Refs. 32, 34, 35, and 36, where the solution is
expressed in terms of "self-added" and "added" mass
coefficients. The self-added mass coefficients characterize the hydrodynamic forces on a member from
its own motion with all other members held
stationary. The added mass coefficients characterize

the hydrodynamic forces in a stationary member

with other members in motion. Because potential
theory is linear, reciprocity applies; i.e., the force
induced onto member i from the motion of memberj
is the same as the force induced o.ito memberj from
the motion of member i.
Experimental comparisons with theory are given in
Refs. 30, 32, and 36 for a seven-member hexagonal
array (Fig. 20) and a 3 x 3 square array (Fig. 21), In
both configurations, the central, member is in motion
while the rest are stationary. The self-added mass
coefficients for the central member are determined
for four sizes of space between members. The
Comparisons between theory and experiment are
made for a row of five cantilevered cylinders, a group
of three cantilevered cylinders, and a group of four
cantilevered cylinders48 in terms of natural frequencies and mode shapes. The group arrangements are
shown in Figs. 22, 23, 24, and 25, respectively, and
the comparisons in Figs. 26,27, and 28, respectively.
The agreement between theory and experiment is
good. Further comparisons were made in terms of
acceleration response under steady-state sinusoidal
excitation for the row of five cylinders and the group
of three cylinders. The frequency of excitation was
swept from 50 Hz to 80 Hz. The comparisons are
shown in Figs. 29 and 30 for the two cases,
respectively. The agreement is good in Fig. 29 and
fair in Fig. 30.

4.0

|or(Gv1.5Gx)

|Gyt@@
1

O -^ 2FTR5) (T) (4)

G

||~0||<G y =G x )J
I
I
I
I
1 L
0.4
0.8
1.2
1.6
Gap-radius ratio Gx/R

== Theory
o Experiment {Moretti & Lowery)
Fig. 21. Theoretical and experimental values of added mass
coefficients for a nine-rod bundle."

00000
G|
G

ii'r

-^ = 2.0,1.0,0.2
3.0

Fig. 22.

CD

O
u
(/>

Row of five cantilevered cylinders.41

15)(6J
2.0

c/i
CO

y
o
0}
-a 1.0
<

Theory
o Experiment
(Moretti & Lowery)
i

,1

0.4
0.8
1.2
Gap-radius ratio Q/R

1.6
-=2.0,1.0,0.5
R

Fig. 20. Theoretical and experimental values of added mass

coefficients for a seven-rod bundle.31

Fig. 23. Group of three cantilevered cylinders.**'/

33

high range of displacement amplitudes applicable for

single isolated members; see Section 5.7.
The added mass and lift forces for a 4 x 4 array of
cylinders moving in unison, as shown in Fig. 31, are
calculated using potential theory.10 This illustrates
an important effect resulting from hydrodynamic
coupling among groups of members. The added mass
forces are not necessarily in line with the direction of
the motion, and lift forces tending to act perpendicular to the direction of motion are generated. In
addition, the distribution of forces is not uniform
among the members. The total load on each member,
the vector sum of the added mass and lift forces,

accentuates this nonuniiormtty. in icrmsofutc una!

= 2.0,1-0,0.5

Fig. 24. Array of seven canlilevred cylinders.4*

Fig. 25. A 2 X 2 array of cantilevered cylinders.4'

Although experimental confirmations are few,

they arc generally good for arrays of cylinders.
Combining this with, the excellent confirmation
established for single isolated members that was
discussed earlier in this report, we feel rather
confident that the potential theory will adequately
describe the added mass and lift forces for groups of
cylinders. We would expect the range of applicability
with respect to motion amplitude to be less than that
for single isolated members because of the close
proximity of the members. Whether or not the
potential theory will be adequate under excitations
of normally expected earthquakes is unknown.
However, for the time being, we believe the potential
theory can be assumed adequate based on the rather
34

load, cylinders 4 and 6 in Fig. 31 carry the highest,

while cylinders 1 and 13 carry the lowest. Nonuniformity in the load for arrays are not accounted
for among the design methods in current use outlined
in Section 4 of this report. It may, or may not, be
important, depending on the purpose of the analysis
and on the configuration of the array; however, its
existance and possible effects should be kept in mind.
For the 4 x 4 array shown in Fig. 31, the net lift
force for the entire array is zero because of symmetry.
The total added mass force, however, depends on the
space between members. For the case where ihe X
and Y center-to-center, distance-to-diameter (X/D
and Y/ D) ratios are both 1.5, the total added mass
force is equal to 16 times that of a single isolated
cylinder.i However, for X/D = 1.5 Y/D = 3.0, the
total is 20 times that of a single isolated cylinder.I0 In
the latter case, the total force carried by the array is
greater than that carried by 16 individual isolated
cylinders. The direction of the motion or flow, as well
as the spacing, affect the force magnitudes as shown
in Figs. 32 through 35 taken from Ref. 10. These
effects, as well as the nonuniform load distribution,
underscore the importance of considering hydrodynamic coupling for groups of cylinders.
If the spacing between members is increased, when
will the members respond as if isolated? For a single
isolated circular cylinder the added mass coefficient
is unity, and the lift force is zero. Using these values
as the criteria for defining when members of a group
become essentially isolated, Figs. 33a and 34a
indicate that the members of the 4 x 4 array of Fig. 31
become essentially isolated at X/ D and Y/ D ratios
of 2.5.10 Similarly, an X/ D ratio of 2.5 was obtained
for the case of two parallel cylinders as reported in
Refs.18, 19, and 26. Other arrangements also gave
ratio values of 2.5; these include in-line and staggered
arrangements of three cylinders and arrays other
than 4 x 4 square arrangements (Refs. 18,19,26, 30,
31, 32, 34, 35, 36, and 42). Thus, a ratio value of 2.5
seems to be more or less universally applicable.

Experiment

Theory

1.

57.30 Hz

56.45 Hz

2.

64.65 Hz

64.56 Hz

3.

69.14 Hz

69.67 Hz

4.

72.45 Hz

73.79 Hz

5.

75.39 Hz

I'

75.65 Hz

Fig. 26. Mode shapes of a row of Ave tubes with a G/R = 0.25."
35

Experiment

Theory

57.52 Hz

58.15 Hz

62.30 Hz

62.32 Hz
\

64.84 Hz

64.86 Hz
/

70.22 Hz

70.14 Hz

Z?

/
72.66 Hz

72.85 Hz
^

74.32 Hz

74.51 Hz

Fig. 27. Mode shapes of a group of three lubes wilh a G/R = O.5.4*

Experiment

1.

^
56.79 Hz

2.

62.79 H z /

Theory

Experiment

/
57.90 Hz

5.

^
/
69.53 Hz

*
69.66 Hz

6.

71.88 Hz

71.38 Hz

62.99 Hz
/

vV'

3. 66.26 Hz ^ \ \

4.

67.38 Hz

Theory

66.73 H z /

67.87 Hz

Fig. 28.

7.

74.32 Hz

74.19 Hz

8.

76.95 Hz

76.39 Hz

v /

Mode shapes of a group of four lubes in unconfined wvtcr.4"

For applications practical to large arrays, a

simplification was suggested in Ref. 35. In a regular
array of cylinders the most significant hydrodynamic
coupling for a given member is with its immediate
neighbors, and coupling with members further
removed can be neglected. This means that,
regardless of the size of the array, it can be analyzed
in subparts consisting of each member and its
immediate neighbors. The author of Ref. 35 reached
this conclusion upon theoretically analyzing regular
hexagonal arrays of 7,19, and 37 cylinders. For each
case, the added mass coefficients of the central
cylinder was calculated and compared from case to
case. The coefficients for the 19- and 37-cylinder
arrays matched almost identically, and those for the
7-cylinder array were close. Thus, the simplification
suggested seemed reasonable. We can make a further
confirmation of this simplification by examining the
4 x 4 array shown in Fig. 31. Considering the high

degree of similarity in the magnitudes and directions

of the forces on cylinders 6, 7, 10, and II, each of
these four centrally located cylinders must be
influenced to approximately the same degree by the
hydrodynamic coupling. Each is surrounded in the
same manner by eight immediate neighbors so, had
the simplification been applied, the forces on each
would have been the same. This comparison may not
provide strong additional confirmation, but it is
supportive. A precaution needs to be mentioned; the
highest loads are carried by the corner members,
cylinders 14 and 16. The simplification was
developed based on results for centrally located
members, so that it may, or may not, apply to
peripheral and corner members. In the case of the
4 x 4 array, the corner members are the most important to analyze. We suspect the corner members
might also be the most important to analyze in other
size arrays.
37

Theory

Experiment
4

2 -

Tube 1

A A A

"1

00Excitation

50

55

60
65
70
75
Forcing frequency Hz

80

50

55

60
65
70
75
Forcing frequency Hz

Fig. 29. Steady-state responses of a row of five tubes to an excitation on tube 5 with a G/R - 1.0."

38

80

Experiment
4

Theory

Tubel
2

x direction

iiir

4F^

FT

Tube2
y direction

0
6
c
o

Av

JU

Tube 3
x direction

03

L.
_QJ
CD
O
U

2 -

to
0)
-Q
3

0'

4 r- -

1
Tube 1

y direction / I

l\

j^rl V ^ U . .

ir

55

SO. 65

70

75

Forcing frequency Hz

80

50

55

60

65

70

75

80

Forcing frequency Hz

Fig. 30. Steady-state response* o f a group of three tubes to an excitation on tube 3 with a G / R = 2.0."

39

f Added-mass forces
- Convective-forces

14

15

10

11

16

Direction of
fluid flow
(for wave)

Direction of
cylinder motion
(for earthquakes)

-X-

Fiji. .11. Added-mass and convective forces on a 4 X 4 array with X = Y = 1.5 D. 10

An approximation that is even simpler was

suggested in Ref. 10. The total interaction is
considered to be the sum of interactions between
each two adjacent members. Therefore, the members
of the array are analyzed two at a time and the results
superimposed. The approximation was shown to be
accurate to within 2 to 25% for a 4 x 4 array
compared to a rigorous analysis. The accuracy varied
depending on the member of the array. Corner
members can be analyzed.
6.3

Hydrodynamic Coupling for Rigid

Members Surrounded by a Rigid
Circular Cylinder

A number of different member arrangements

surrounded by a rigid circular cylinder have been
40

investigated. Closed-form solutions based on

potential theory are presented for coaxial cylinders
(Refs. 7,30,26, and 34), eccentric cylinders (Refs. 26
and 34), and an array of cylinders surrounded by a
cylinder (Refs. 26 and 34). A finite element method
was developed and applied to coaxial cylinders and
to an array of cylinders surrounded by a circular
cylinder.21-42 Analyses of coaxial cylinders using an
incompressible viscous fluid theory are given in Refs.
27 and 45. For the remainder of Section 6.3 of this
report, we will focus primarily on results pertaining
to rigid coaxial cylinders. This is a simple model
commonly used to simulate the internals of the
reactor vessel under seismic excitation.
For two rigid coaxial cylinders in motion, as
shown in Fig. 36, with the annular space filled with
fluid, the potential theory solution is expressible in a
form quite convenient for design applications. The

-l

1.5

1
r
Cylinder No. 1-

X=Y=1.5D

0.5
(a) Added-mass force
_i

i_

60

-J_

120
Angle of attack deg

Fig. 32.

Hjdrodynamlc forces vs angle of attack for a 4 X "* array with X = Y = 1.5 D. 10

JL.

180

(D
XJ
=J
+->

'E
en
(O

0.5(a) Added-mass force

D
CD
XJ

01-

8-S

5-v

O .

* "** ^ /

V
c
CD Ol
>

to
u

0.4

(b) Convective-force

o 8

O S

% - ^_

*"""* " ^ ^

'

180

60

120
Angle of attack - deg

Fig. 33.

42

Hydrodynamic forces vs angle of attack for a 4 / 4 array with X = 1.5 D and Y = 3 D . "

Y/X = 2, a = -45
Added-mass force
U

Convective force
Cylinder No. 1

(a) Force coefficients

-ff-

X/D

Fig. 34. Hydrodynamic forces vs spacing for a 4 X 4 array with Y/X = 1.0 and a 45 deg.'

Y/X= 1,a= 90
Cylinder

~"

Added-mass force

""

Convective force
Approximate method

Added-mass force on cylinder No. 2

Convective-force on cylinder No. 1

(a) Force coefficients

>
1351ifr
2,5
/ "

90

TotalV

45 "
(b) Direction of force

_ _ ^ _

X.
; /
-45

-ffX/D

Fig. 35.

Hydrodynamic forces vs spacing for a 4 X 4 array with Y/X = 2.0 and a = 90 deg.10

Inner cylindrical '

(or spherical)
body

^Outer cylindrical
(or spherical)
surface

Fig. 36. Two-body motion with fluid coupling.7

fluid forces on the inner and outer cylinders are,

respectively,7
Fr, = -M H Xi + (M, MH)& ,
F f 2 = (M, + MH)Xi - (Mi

(1)
M2 + M H )x 2

,(2)

Mi = na2Lp = mass of fluid displaced

by the inner cylinder
M2 = 7rb2 Lp = mass of fluid that could
fill the outer cylindrical
cavity in the absence of
the inner cylinder

(3)

where,

- mass term depending on the

relative sizes of the inner and
outer cylinders.
L = length of the cylinders
P = mass density of the fluid.

(4)

(5)

Values for M, and M, can be determined experimentally or theoretically. Similar expressions are
also presented in Ref. 30. These equations theoretically apply only to infinitely long cylinders;
therefore, L should be significantly greater than the
radii a and b. In addition, we expect that the
solution's invalidity will diminish if thi. annular space
is very small compared to radii a or b because the
fluid would then be subjected to a significant amount
of flow and shearing to accommodate the relative
motions of the cylinders. The incompressible and
invicid assumptions would be less valid. Unfortunately, we have found no published indication of the
range of applicability of the Eqs. (1) through (5) with
respect to annulus size and motion amplitude.
Some comparison with experiments for five cases
of two coaxial rigid cylinders are given in Table 9
taken from Ref. 7. The outer cylinder is fixed while
the inner cylinder is vibrated. The added mass values
on the inner cylinder, evaluated with Eq. (1), were
compared with measured values. In the first four
cases, the theoretical value was higher than the
experimental by 21 to 36%, and, in the fifth case, it
was lower by 33%. The comparison was fair.
The finite element technique developed in Refs. 21
and 42 compared very well with potential theory in
terms of added mass coefficients for two coaxial rigid
cylinders. (See Table 10 from Ref. 21). The basis of
comparison was the Mi, M2, and M4 of Eqs. (1) and
(2). Therefore, the finite element technique is capable
of duplicating the closed-form results very well. A
comparison of the finite element technique with experimental results was presented in Ref. 42 for a 2 x 2
array of square cylinders surrounded by a circular
cylinder, Fig. 37. Cylinder B is driven at a fixed
displacement amplitude over the frequency range
from 3 to 15 Hz, and the required force was
monitored. The agreement between the finiteelement and experimental results was reasonable. 4 '
A somewhat more sophisticated treatment of
coaxial rigid cylinders is given in Refs. 27 and 45
using an incompressible viscous theory. The solution
expressions are much more complex than those for
potential theory and are contained in the references.
A comparison with experiment was made for a fixed
outer cylinder and oscillating inner cylinder. The
outer cylinder diameter was varied from 0.625 in. to
2.5 in., while the inner cylinder diameter was kept at
0.5 in. The agreement between analysis and
experiment was quite good, as shown in Fig. 38, and
it is noticeably better than the comparisons discussed
earlier for the potential theory. A possible conclusion
is that viscous effects may be important and perhaps
should be included when analyzing coaxial rigid
cylinders. More experimental comparisons are
needed to confirm this possibility.
45

Table 9. Added mass on coaxial rigid cylinders,7

Annulus
Radius,
in.

Clearance,
in.

4.0
3.9
4.0
4.0

0.16
0.39
0.25
0.25

4.0

0.25

Liquid
Water
Water
Water
Glycerol
solution
Oil

Nalural
frequency,
cpm

Calculated
added mass,
lbs

Experimental
added mass,
lbs

370
520
425
390

280
100
170
190

180
75
127
150

27
25
21

320

150

200

-33

Difference,

%
36

Table 10. Comparing added mass coefficients between closed form and finite elements solutions for coaxial
rigid cylinders. 4 2

MH

= /OTa

M, + M2 + M H = pmV

Mi + MH =

-2pn

mr

U2-a2)

Closed form

Finite clement

5.23 p

5.169 p

20.95 p

20.792 p

-8.38 p

-8.29

*a = radius of inner cylinder = 1 in.

b = radius of outer cylinder - 2 in.

3.75
.O J2

i.'e '
- I 2.50

Ig
E o
CO

Qj

Beam B
Prediction
Experiment

51.1.25

10
12
14
Forcing frequency - Hz .

16

Fig. 37.;' Displacement response of a 2 X 2 square cylinder surrounded by a circular cylinder/

18

6.4

Hydrodynamic Coupling for Flexible

Coaxial Cylinders

Coaxial cylinders with the inner cylinder analyzed

as a flexible shell probably constitute a more realistic
model of the internals of the reactor vessel than
would coaxial rigid cylinders. Such a mathematical
model was analyzed in Ref. 38 using an incompressible invicid theory. The deformation of the
inner cylinder is compared with experiment in Fig. 39
taken from Ref. 38, and the comparison is
reasonable.

influenced by member arrangement, spacing, and

relative motions among the members.
Analyses were carried out for two coaxial
cylinders using a viscous fluid theory.-'7 '^ Three
fluids were investigated in Ref. 27, and the
theoretical damping was compared with the experimental as shown in Fig. 39 taken from Rcf. 27. The
agreement was quite good, indicating it is possible to
obtain reliable damping values theoretically. Agreement was not as good in Ref. 45, where, by
comparing the theoretical and experimental oscillatory motion amplitudes, it was determined that the

I he case of three coaxial eyiiinaeis wii.Ii ilic outer

theoretical damping uiiueiesi.iuiai.cu tlic actual by U

cylinder rigid and the inner cylinders flexible was

analyzed in Ref. 46 using a compressible invicid fluid
theory. A simpler case involving only one inner
cylinder was compared with experiment in terms of
natural frequencies in Tables 11 and 12 taken from
Ref. 46. The m and n quantities are, respectively, the
axial and circumferential mode numbers of the inner
cylinder. The agreement between theory and experiment is very good.
A finite element analysis using the code N ASTR AN
was applied to two coaxial cylinders, the outer one
rigid and the inner flexible.''7 A compressible invicid
fluid theory was used. A comparison between
analytical natural frequencies and experimental data
is shown in Fig. 40 taken from Ref. 47, The
comparison ranged from good to fair.
In general, based on the very limited amount of
experimental comparison, the compressible invicid
fluid theory seems to do better than the incompressible invicid potential theory. This indicates that
fluid compressibility may be quite important to
include when analyzing flexible members. Additional experimental confirmation is needed to fully
establish this possibility.

mcasureable amount.
The dependence of damping on the size of the
annular space between two coaxial cylinders is
clearly seen in Fig. 41. A sharp increase in damping is
seen at a D/d ratio less than 1.75 to 2.75, depending
on the fluid involved; the value of 1.75 applies to
water. The quantities D and d are the diameters of
the outer and inner cylinders, respectively. The
diameter d was 0.5 in., and D varied from 0.625 in. to
2.5 in.
Experimental damping values for coaxial rigid
cylinders submerged in three fluids are shown in
Table 9 taken from Ref. 7. Adding the values for
water to Fig. 41 indicates good agreement with the
data from Ref. 27.
Experimentally determined damping from water
viscosity are presented in Ref. 48 for a row of fivecylinders (Fig. 22), a group of three cylinders (Fig.
23), a hexagonal array of seven cylinders (Fig. 24). a
2 x 2 array of cylinders (Fig. 25). a 2 x 2 array of
cylinders near a wall (Fig, 42), and a 2 x 2 array of
cylinders surrounded by a cylinder (Fig. 43). The
results from Ref. 48 are reproduced in Tables 13
through 25. The tubes are all 0.5 in. diameter and
12.0 in. long. The damping values in these tables
should be approximately the same as those in Fig. 41
because the inner cylinder used for Fig. 41 was also
0.5 in. diameter, and the space between cylinders
reported in Tables 13 through 25 are generally within
the gap size range covered in Fig. 41. In other words,
space size-to-radius ratio values of 0.4 to 2.0 for Ref.
48 corresponds to D/d ratio values of 1.8 to 5.0 for
Ref. 27. Comparing the damping values confirmed,
our speculation; i.e., the added damping values from
Tables 13 through 25 ranged from 0.38 to 1.9%; this
range compares very closely with the range 0.5 to
1.8% shown for water in Fig. 41 and corresponding
to D/d ratios from 1.8 to 5.0.
Up to this point, all experimental data for
damping are mutually supportive, and the damping
for multiple members 0.5-in. in diameter is characterized to a usable degree. The next question is how can

6.5

Damping for Multiple Members

In Section 5.5 of this report we explained that for

fully submerged structures in a finite size container,
radiation damping can generally be ignored. The
contributions to added damping that remain are
fluid viscosity and component impact. Both theoretical and experimental values for fluid viscosity
damping have been published, although no analytical treatment of impact damping has been found. For
experiments involving both fluid viscosity and
component impact, no separation of the measured
total damping into these two contributions was
made. Establishing a fixed value of damping for a
general multiple member structure is very difficult, if
not impossible, because damping can be significantly

47

7 -

Test
No.

Fluid

Water

<fn>Air

16.04
4

PS

Mineral oil
Mineral oil

16.04
16.43 "

3
.3

Silicone oil

16.04

Silicone oil

16.26 *

Water

58.38

Theory
Calculated values
o

Test No.
3

D/d

Fig. 38.

48

Theoretical and experimental values of Cm as a function of D/d. 2 7

Zero displacement line

1 mil
(0.0254 mm)

Wall thickness = 1/8

Material = stainless steel
Axial location = 18

Computed worst possible displacement (all ring modes in phase)

0.43-1.34 mils (0.0109 - 0.0340 mm)

Computed host probable displacement (random phase between

ring modes) 0.94 miis (0.0239 mm)
O Measured displacement 0.90 - 1.28 mils (0.0229 - 0.0320 mm)

Fig. 39.

Computed vs r.m.s. displacement of cylinder."

49

Table II. Measured and computed natural frequencies for coaxial cylinders, inner cylinder flexible, inner
cylinder filled with water only.42
rtyd<\
~i V
I

V. 5

<;)

Experimental
frequency, 11/

Computed
frequency. H i

Discrepancies %
frequency

10.1

90.6

12.0

106

98.2

7.4
0.7
0.5
7.3

145
213
219

144
214
203

Table 12. Measured and computed natural frequencies for coaxial cylinders, inner cylinder flexible, both
inner and outer cylinders filled with water."
Mode
(m. n)

1.2
1, 3

1,4
1.5
1.6
1,7
2,5
2.6

Experimental frequency,

11/
64
54
60
90
139
206
136
165

we extrapolate the results to structural sizes of

concern. At this point, there is no established way.
The extrapolation technique we developed for single
isolated members, (Figs. Hand 15)cannot be trusted
to apply to multiple members in the absence of
experimental evidence. A possible method is the
analytical technique developed in Ref. 27 to form
relationships that can relate the 0.5-in. specimens to
structural sizes of concern. This might be a very
useful area for future exploration because damping is
a topic of high interest.
Some measurements of total damping in actual
reactors and models of reactors are reported in Ref.
41. The values arc 2 to 5% for core-barrel beam
modes, 1 to 2% for core-barrel shell modes, 2 to 5%
for guide tubes. These values are measured under
low-displacement amplitudes on actual reactors.
When the coolant is flowing, the damping increases
with increasing flow rate, giving riss to core barrel
damping ranging from 8.8 to 12%. In the opinion of
the authors of Ref. 41, a significant contribution to

50

Computed frequency,
Hz

Discrepancies, %
frequency

62.4

2.5

46.2

14.0

55.9

6.8
0.9
1.4
1.6
6.3
0.8

89.2
140.9
209.2
127.4
163.7

the total damping resulted from component impact,

particularly while the coolant was flowing. Component impact was, therefore, very possibly responsible for a major part of the 8.8 to 12% damping
measured. Unfortunately, no separation between
fluid viscosity effects and component impact was
made. Consequently, the usefulness of the damping
values is limited for general application because the
component impact contribution could vary from one
reactor design to another.
Further evidence that component impact contributes significantly to the damping was found in
Ref. 43. The effect of tube-support interaction on the
dynamic response of heat-exchanger tubes was
examined. The total damping measured was from
2 to 7.5%. whereas it was felt the combined
structural damping and fluid viscosity damping
should have been approximately 2%. Again, no
separation between component impact and other
contributions was made.

iir
300 -

Experimental data
Extrapolation from
in-air data
NASTRAN predictions

250

200
IM

X
u
c

n_ rn
8 2

0)
3
LL.

(6 2)
7 2
100

6 2
(6 1)
6 1
50

(5 1)
5 1
4 1

I I

0.1

0.2

0.4

0.6

0.8

1.0

Gap size in.

Fig. 40.

Comparison of experimental and predicted vibration frequencies for the shell with a fluid filled gap. 4 '

20

18

Test No.

Fluid

(f

Water

16.04

fvmitMdi o i l

V
A

3
4

Silicone oil
Water

n>Air

16.04
68.38

Theory

From Ref. 7

Test No.

Fig. 41. Theoretical and experimental values of C as a function of D/d.3

52

y////////////////////////////////

CJ

2R
G
Gw/R = 0.5,1.0
G/R = 0, 0.4, 0.8
Rc/R = 4.0, 3.5

Fig. 42. A 2 X 2 army of caiiiilevcrid cylinders near a wall."

Hg. 43. A 2 x 2 array uf cautikviieii cylinders near a wall."

Table 13. Experimental and analytical results for uncoupled vibration of a row of five tubes.48
Cap-toradius Direction
ratio,
of
Tube
C/R
motion number

2.0
(1.988)

1.0
(0.981)

0.25
(0.248)

Dimensionless
spring
constant,

Measured
uncoupled natural
frequency, Hz

Measured
damping
ratio

In air

In water

In air

In water

Calculated
uncoupled natural
frequency in
water, Hz

104 J

83.79

69.63

0.00118

0.0044

70.07

99.0

84.67

70.41

0.00095

0.0049

70.58

128.5

84.27

69.92

0.00031

0.0053

70.39

I5S.2

84.38

70.21

0.00028

0.0045

70.57

354.0

77.93

65.50

0.00062

0.0043

66.87

129.0

84.08

70.02

0.00113

0.0042

70.31

129.3

85.05

70.90

0.00032

0.0036

70.89

242.6

84.86

70.51

0.00103

0.0044

70.86

288.5

84.86

70.80

0.00078

0.0O34

70.96

354.0

77.93

66.70

0.00148

0.0042

66.87

105.1

83.40

69.24

0.00044

0.0063

69.54

84.2

83.79

68.75

0.00052

0.0078

69.39

98.3

83.69

68.85

0.00051

0.0088

69.43

86.5

83.01

68.26

0.00063

0.0099

68.96

186.6

77.05

64.94

0.00103

0.0047

65.93

155.8

83.89

69.92

0.00032

0.0037

69.91

114.1

84.28

69.53

0.00054

0.0043

69.72

154.9

84.28

69.63

0.00073

0.0044

69.82

127.4

83.59

69.04

0.00152

0.0051

69.37

253.4

77.25

65.53

0.00092

0.0039

66.07

99.2

83.40

6855

0.00093

0.0073

68.45

85.2

83.50

66.99

0.00045

0.0124

67.08

92.0

83.50

67.29

0.00146

0.0137

67.23

97.2

83.40

67.38

0.00131

0.0190

67.28

325.2

77.73

6455

0.00108

0.0081

65.63

131.5

83.79

69.24

0.00046

0.0062

6852 '

101.1

83.79

67.38

0.00055

0.0076

66.70

97.2

83.59

66.80

0.00088

0.0070

6653

103.6

83.50

67.19

0.00069

0.0076

66.75

930.5

78.02

65.42

0.00226

0.0046

65.66

53

Table 14.

Experimental and analytical results for uncoupled vibration of a group of three tubes.4"

CJaptoradius Direction
ratio.
of
Tube
motion
number
G/R
1
X

2.0
ll.v.io)

Dimcnsionless
spring
constant,

Measured
uncoupled natural
frequency, \\i.

Measured
damping
ratio

Calculated
uncoupled natural
frequency in
water, Hz

P.

In air

In water

In air

In water

53.2

82.03

68.35

0.00152

0.0038
0.G04I

70.35

68.53

141.4

84.38

70.21

0.00090

3X6.0

77.24

65.82

0.00341

0.0041

66.22

t t >

K? M

(<H.*,S

0.(10103

0.0037

68.59

1.0
(0.983)
y

0.5
(0.475)
y

79.1

83.50

69.53

0.00095

0.0045

6'J.63

123.2

76.46

65.33

O.002I7

0.0034

65.55

62.6

82.23

69.26

0.00072

0.0047

68.28

79.5

83.50

69.14

0.00109

0.0087

69.08

115.8

76.76

65.04

0.00073

0.0071

65.36

71.1

82.52

68.36

0.00095

0.0042

68.37

63.4

83.01

68.55

0.00174

0.0075

68.75

77.4

76.17

64.55

0.00215

0.0063

64.91

58.5

82.42

67.58

0.00076

0.0044

67.62

94.0

83.98

67.68

0.00183

0.0051

68.26

82.9

76.07

63.09

0.00226

0.0052

63.76

52.6

82.13

66.99

0.00125

0.0048

66.75

59.2

83.01

67.77

0.00233

0.0053

67.79

88.3

76.17

63.87

0.00119

0.0055

64.12

ill!

Table 15.

G/R

Experimental and analytical rciiilts fur uncoupled vibration of a group of seven tubes.'

Direction
Tube
of
motion number

1.0
(0.867)

0.4
(0.394)

Pi

Measured
uncoupled natural
frequency, 11/
In air

In water

Measured
damping
ratio
In air

In water

Calculated
uncoupled natural
frequency in
water. 11/

52.2

82.42

67.91

0.00079

0.0049

68.29

60.1

83.11

68.96

0.00044

0.0103

68.72

73.6

83.06

68.75

0.00086

0.0043

69.94

57.3
; o

81.93

67.48

0.00050

<;
1.5
(1.384)

Dimensionless
spring
constant.

oi no

0.0061

67.97

(i.Oli.'-i

oti.Oj

223.4

77.15

65.38

0.00052

0.0040

65.77

47.4

81.78

67.19

0.00099

0.0047

67.29

68.80

0.00107

64.1

82.96

0.0040

68.81

79.5

83.74

69.17

0.00089

0.0045

69.32

55.2

82.37

67.63

0.00140

0.0055

68.21

72.0

82.47

68.46

0.00110

0.0033

68.49

73.3

82.62

68.75

0.00054

0.0043

68.64

106.4

76.46

64.45

0.00094

0.0039

65 05

49.U

81.93

66.80

0.00109

0.0047

67.41

52.1

81.15

65.63

0.00104

0.0049

66.20
66.61

54.7

81.84

66.21

0.00063

0.0040

91.9

82.81

67.77

0.00083

0.0043

67.96

75.7

82.62

67.19

0.00084

0.0040

67.48

62.0

82.23

67.38

0.00085

0.0042

67.20

88.3

75.49

63.39

0.00180

O.O043

63.68

50.6

81.35

64.84

0.00092

0.005J

65.18

71.5

81.93

67.19

0.00061

0.0044

67.09

72.5

82.52

67.19

0.00063

0.0055

67.42

61.2

81.98

66.60

0.00062

0.0049

66.75

75.7

82.62

67.97

0.00084

0.0040

67.74

63.2

82.28

68.16

0.(10120

0.0041

67.49

74.2

75.20

62.60

0.00134

0.0048

63.00

48.9

81.25

64.84

0.00078

0.0055

65.10
64.21

66.2

82.62

67.20

0.00253

0.0072

66.8

83.06

62.82

0.00263

0.0073

64.27

61.1

82.62

65.14

0.00143

0.0056

65.72

95.9

83.01

63.55

0.00121

.0.0064

64.52

91.2

82.81

61.91

0.00130

0.0010

64.40

76.5

75.98

61.77

0.00048

0.0060

62.39

48.8

81.64

59.13

0.00168

0.0103

60.13

56.6

82.23

65.72

0.00250

0.0052

64.85

53.2

82.62

66.11

0.00200

0.0050

64.99

49.3

82.03

61.87

0.00082

0.0088

63.14

77.1

82.62

69.77

0.00126

0.0050

65.26

91.2

82.81

65.23

0.00118

0.0054

65.45

62.6

75.59

59.52

0.00045

0.0066

60.28

40.8

81.05

58.94

0.00062

0.0112

59.69

Table 16. Experimental and analytical results for uncoupled vibration of the four-tube a m y in
unconfined water.4*
Ciap-toradius Direction
Tube
ralict.
of
mcilion
number
G/R

Dimensionless
spring
constant,
P.

In air

1I n H a i t i

In air

In water

Calculated
uncoupled natural
frequency in
water. Hz

mil

(M70

AR A*

0(10099

0.0079

68.57

Measured
uncoupled natural
frequency, Hz

Measured
damping
ratio

60.6

83.98

68.85

U.U0169

0.0138

68.42

78.1

83.59

68.55

0.00566

0.0125

68.41

0.5

1030.0

77.83

65.33

0.00091

0.0080

65.55

(0.585)

75.0

83.30

67.77

0.00131

0.0072

68.17

Table 17.
Gap-to
radius
ratio,
G./R

74.9

84.28

68.95

0.00286

0.0147

68.82

86.4

83.78

68.46

0.00290

0.0089

68.57

228.1

77.34

64.16

0.00633

0.0117

65.13

Experimental and analytical results for uncoupled vibration of the four-tube array near a flat wall.48
Direction
of
motion

1.0

0.5

56

2
3

Tube
number

Measured
uncoupled natural
frequency. III

Measured
damping
ratio

Calculated
uncoupled natural
frequency, Hz

68.46

0.0066

68.27

67.24

0.0071

66.87

66.75

0.0059

66.88

65.14

0.0101

65.29

67.58

0.0078

67.93

67.04

0.0065

67.38

66.55

0.0087

67.15

64.06

0.0070

64.93

68.51

0.0059

68.14

65.53

0.0138

65.31

65.38

0.0143

65.31

65.19

O.O094

65.18

67.48

0.0038

67.85

66.16

0.0087

66.15

66.06

0.0096

65.91

64.21

0.0103

64.86

Table 18. Experimental and analytical results for uncoupled vibration of the four-tube array confined
in a cylinder.48
Radius
ratio,
R, / R

Eccentricity,
G./R

Direct' /n
of
motion

\
0.0

Tube
number

Measured
uncoupled natural
frequency, Hz

Measured
damping
ratio

Calculated
uncoupled natural
frequency, Hz

65.33

0.0095

66.07

65.77

0.0082

65.89

65.92

0.0071

65.91

62.26

0.0055

63.43

64.70

0.0079

65.68

65.5}

0.007}

M.2H

4.0

0.4

0.8

0.0

3.5

0.4

65.87

0.0068

66.06

62.26

0.0053

63.02
65.47

65.48

0.0060

65.14

0.0070

66.19

65.38

0.0068

66.21

62.60

0.0054

62.92

64.94

0.0069

65.07

65.33

0.0060

6t.62

64.84

0.0106

66.40

62.55

0.0054

62.50

65.33

0.0069

64.25

62.84

0.0093

66.30

63.82

0.0082

66.31

62.79

0.0074

61.87

65.82

0.0057

63.89

64.21

0.0091

66.79

62.94

0.0111

65.57

62.45

0.0075

61.49

63.67

0.0079

64.43

64.31

0.0092

64.25

64.75

0.0094

64.27

60.84

0.0105

62.02

62.21

0.0101

64.05

64.50

0.0091

64.63

63.82

0.0100

64.42

59.91

0.0102

61.62

64.65

0.0069

62.73

63.67

0.0085

64.94

63.92

0.0070

64.96

61.13

0.0128

50.56

63.78

0.0085

62.36

62.45

0.0073

65.40

63.92

0.0090

65.19

61.67

0.0067

60.17

57

Table 19. Experimental and analytical results for coupled vibration of five tubes/'
('ap-fo-radius
ratio,
C./H

Direction
of
motion

Mode
number

Measured coupled
natural
frequencies, Hz

(1.988)

1.0
(0.981)

Calculated coupled
natural
frequencies, H *

Damping
ratio

66.16

66.45

0.0043

68.12

68.50

0.0047

69.34

69.46

0.0047

71.12

71.15

0.0046

73.29

73.28

0.0051

66.0K

66.18

0.004U

68.64

68.44

0.0039

70.21

70.54

0.1040

71.80

72.08

0.0038

72.68

73.04

0.0041

63.77

64.56

0.0061

65.72

66.24

0.0072

68.09

68.16

0.0075

70.98

70.91

0.0080

74.83

74.48

0.0088

63.77

63.35

0.0040

66.67

66.74

0.0039

69.80

69.98

0.0043

71.90

72.31

0.0046

73.52

73.88

0.0046

59.96

6102

0.0126

62.50

62.92

0.0105

66.29

66.59

0.0102

4
0.25
(0.248)
'

71.53

71.55

0.0127

77.64

77.19

0.0152

57.33

56.45

0.0058

64.67

64.56

0.0059

69.14

69.67

0.0063

72.95

73.39

0.0074

75.39

75.65

0.0081

Table 20. Experimental and analytical results for coupled vibration of three tubes.4"
Gap-to-radius
ratio,

Mode

Measured coupled

Calculated coupled

Damping

G/R

number

r.alural frequencies, Hz

natural frequencies Hz

ratio

2.0
(1.933)

1.0
(0.983)

0.5
(0.475)

>

64.02

64.61

0.0036

65.20

65.23

0.0038

67.65

67.76

0.0039

69.04

69.10

0.0039

71.10

0.0(143

71.52

71.61

O.0U42

61.21

62.04

0.0061

64.02

63.84

0.0063

66.25

66.42

0.0060

69.52

69.68

0.0065

72.10

71.78

0.0069

72.85

72.72

0.0070
0.0046

57.42

58.15

62.29

62.32

0.0051

64.76

64.86

0.0047

70.24

70.14

0.0055

72.88

72.66

0.0054

74.47

74.32

0.0055

Table 21. Experimental and analytical results for coupled vibration of seven tubes.4
(iap-lo-radlus
ratio.
0/R

1.5

(1.384)

1.0
(0.867)

0.4
(0.394)

frequencies, Hz

Calculated
natural
frequencies, Hz

Calculated
damping
ratio

60.25

61.18

0.0043

61.91

62.66

0.0045

62.70

62.76

0.0052

Mode
number

Measured
natural

64.65

64.74

0.0040

66.06

66.33

0.0045

bhAb

66.94

0.00-JS

67.53

68.31

0.0043

68.99

68.79

0.0055

69.87

70.22

0.0051

10

70.75

71.07

0.0060

- 11

71.58

72.03

0.0066

12

72.41

73.13

0.0050

13

73.44

73.32

0.0056

14

74.36

74.47

0.0049

55.61

56.87

0.0041

58.20

58.73

0.0044

58.89

58.86

0.0042

62.06

62.12

0.0041

64.45

64.43

0.0043

65.33

64.99

0.0042

68.36

67.68

0.0050

69.29

68.23

0.0044

70.85

70.12

0.0049

10

71.63

71.26

0.0047

11

72.66

72.69

0.0048

12

74.07

74.04

0.0052

13

74.46

74.27

0.0056

14

75.54

75.79

0.0050

48.39

49.69

0.0048

50.96

51.35

0.0077

51.41

51.70

0.0066

59.47

59 JO

0.0052

60.99

61JS

0.0054

62.40

62.09

0.0050

68.41

67.99

0.0067

69.14

68.30

0.0057

70.80

70.74

0.0074

10

72.46

72.09

0.0062

11

74.46

74.22

0.0079

12

76.07

76.02

0.0094

13

76.46

76.09

0.0099

14

78.56

78.28

0.0065

Table 22.
water.4"

Experimental and analytical results for coupled vibration of the four-tube array in unconfined

Gap-to-radius
ratio,
G/R

0.5

Mode
number

Measured coupled
natural
frequencies, Hz

'alculated coupled
natural

Damping

frequencies, 11/

ratio

56.79

57.90

0.0090

62.79

62.99

0.0094

66.26

66.73

0.0123

67.38

67.87

0.009!

U>.Uo

V.<* 1 t I

(U.3f>)

71.88

71.38

0.0095

74.32

74.19

0.0124

76.95

76.39

0.0121

Table 23. Experimental and analytical results for coupled vibration of the four-tube array near a flat wall.48
Gap-to-radius
ratio.
G./R

1.0

0.5

Mode
number

Measured coupled
natural frequency,
Hz

Calculated coupled
natural frequency,
Hz

Damping
ratio

56.69

57.78

0.0066

61.87

62.08

0.0071

64.84

64.75

0.0077

66.94

66.69

0.0071

69.04

68.95

0.0079

71.19

70.32

0.0C77

73.54

73.35

0.0080

76.81

76.35

0.0083

56.25

57.17

0.0094

60.11

60.52

0.0093

64.36

64.08

0.0098

66.70

66.11

0.0091

68.51

68.30

0.0097

70.68

69.60

0.0112

72.99

72.82

0.0084

76.51

76.13

0.0115

61

Table 24.

Conditions

Experimental result!* for uncoupled vibration of the four-tube arrays in viscous fluids.'
Direction
of
motion

In unconfined
fluid

i'

Near a
flat wall
( G . / R =0.5)
y

7.

Measured uncoupled
natural frequtnc}, 11/
Tube
number

Hater

1
2

Mineral oil

Water

Mineral oil

68.65

68.3!

0.0079

0.0262

68.85

69.14

0.0138

0.0318

68.55

68.41

0.0125

0.02X5

65.33

6S.04

0.0080

0.0309

17.77

67.48

0.0072

0.0284

68.95

69.14

0.0147

0.0336

68.46

68.56

0.0089

0.0285

64.16

64.99

0.0117

0.0290

68.51

68.56

0.0059

0.0254

65.53

66.31

0.0138

0.0311

65.38

65.58

0.0143

0.0308

65.19

64.70

0.0094

0.0375

67.48

67.48

0.0038

0.0350

66.16

65.82

0.0087

0.0268

66.06

66.02

0.0006

0.0323

64.21

64.84

0.0103

0.0243

CONSERVATIVE CHOICE FOR ADDED MASS

The calculation of added mass will generally

involve varying degrees of engineering judgment
regarding such considerations as the effects of finite
length, neighboring members, irregularities in geometry, etc. Decisions on these factors can be
significantly influenced by considering whether
conservatism is increased by maximizingor minimizing the added mass. This generally varies from
situation to situation, and for some cases, a
preliminary analysis may be required to help make
the decision. We will make some suggestions
regarding the structures of concern subjected to
seismic excitations prescribed by the response
spectrum in R.G. I.6050 (shown in Fig. I).
Let us assume that we are interested in the inertial
forces; therefore, we will deal with the spectral
accelerations. The maximum spectral acceleration
occurs at a frequency of 2.5 Hz, as indicated in Fig. 1.
Therefore, to help ensure conservative inertial forces
at frequencies above 2.5 Hz, we would maximize the
added mass to bring the calculated natural frequency
into regions of higher spectral acceleration. Conversely, we would minimize the added- mass at
frequencies below 2.5 Hz to achieve the same
objective.
62

Menured daimping
ratio

Figure I shows the natural frequency ranges of the

structures of concern listed in Table 3. The natural
frequencies are for representative existing structures.
According to the frequency values shown in Fig. 1,
we should maximize the added mass for the reactor
core barrel. For spent-fuel storage racks, the natural
frequency can fall on either side of 2.5 Hz. Therefore,
the natural frequency in air should be first
determined to see whether the added mass should be
maximized or minimized. The natural frequency of
one fuel bundle we examined was close to 2.5 Hz.
Other fuel bundles can presumably fall above or
belcw 2.5 Hz. Therefore, again the natural frequency
In air should be first evaluated to determine whether
to maximize or minimize the added mass. The
natural frequency of the main steam-relief valve line
generally falls above 2.5 Hz; however, the upper limit
is quite close to 2.5 Hz. Therefore, for cases suspected
of having a high natural frequency, it would be best
to check the natural frequency first. Otherwise, we
would generally minimize the added mass for main
steam-relief valve lines. This discussion applies in
principle to all acceleration response spectrums; the
frequency at which the maximum spectral acceleration occurs may differ from the 2.5 Hz applying to
the spectrum of R.G. 1.60.

Table 25. Experimental and analytical results for coupled vibration of the four-tube arra.v contained
in a cylinder.48
Radius
ratio,
R,/H

Eccentricity,

0.0

4.0

0.4

0.8

0.0

3.S

0.4

Mode
number

Measured coupled
nalural frequency,
ill

Calculated coupled
natural frequency,
11/

Dsmping
ratio

56.15

57.86

0.0061

59.52

59.98

0.0058

67.21

62.11

0.0074

t T TT

p "070

67.53

68.31

0.0070

69.43

69.27

0.0080

72.27

72.22

0.0083

76.66

76.43

0.0086

56.25

57.81

0.0058

59.40

59.12

0.0054

61.67

61.96

0.0062

62.26

62.80

0.0075

67.68

68.06

0.0066

69.43

69.15

0.0073

.7

72.11

72.M

0.0079

76.61

76.43

0.0081

55.86

57.29

0.0064

58.11

57.63

0.0071

60.74

61.25

0.0075

62.65

63.14

0.0092

67.29

67.33

0.0076

68.60

68.79

0.0084

71.53

72.34

0.0100

76.32

76.35

0.0099

5S.57

57.43

0.0091

57.03

57.70

0.0090

59.33

59.56

0.0085

59.77

59.83

0.0087

66.31

67.40

0.0105

68.07

68.25

0.0100

70.90

70.59

0.0102

76.22

76.18

0.0113

55.76

55.33

0.0076

57.37

57.23

0.0084

59.20

59.03

0.0078

59.72

60.87

0.0076

66.70

66.42

0.0088

68.21

67.91

0.0087

70.31

70.99

0.0089

76.17

76.12

0.0099

8. COMPUTER CODES IN CURRENT USE

attributes or capabilities f these codes were not
explored for this >.iui!y.

The computer codes used for various aspects of

lluid-structure interaction analysis are listed in Table
26; this is not a complete list of such codes. The

Table. 26. Codes used in industry

linn

Application

Codes
i:t)A(7M.SAfM
ANSVS
SOI.ASURF
SOI.AICE
SOI.AFI.F.X
WECAN
NASTRAN
MARC
DYNA-3D
PISCES
WATERHASS
MULTIFLEX
AMASS
CESIIOCK

HuM motion in tnks

M>At

Offshore reactor platform

Incompressible fluid motions: waves
Compressible fluid motions
Internals of reactor vessrl
possible, said nus.
Internals of reactor vessel
Boaster tanks space shuttle tanks
(eneral, including fluids
General, including fluids
Fluid-structure interaction
Added mass

Offshore I'ower Systems

I.ASI.
I.AS1.

Internals of reactor vessel

Added mass
Fluid-structure interaction

Westinghouse
Argonne National l,ab.
Combustion Engineering

LAS1.
Westinghouse
Universal Analysis
Marc Analysis Corp.
IAA.
Physics International
(,t

9. CONCLUSIONS AND RECOMMENDATIONS

9.1

Idealized Single Isolated, and

Multiple, Members

Hydrodynamic effects on submerged single isolated members arc fairly well understood. The added
mass and added damping concept is adequate
under seismic and normal steam-relief excitations:
however, it is probably inappropriate for blowdown
accidents. The potential theory will accurately give
the added mass values, and tabulated results are
available in the literature for a wide variety of single
member geometries (Table 6). A presentation of the
potential theory can be found in standard textbooks
on the. mechanics of fluids, such as Ref. 16.
Values for added damping are generally determined experimentally, and values are published for
single isolated specimens of small sizes; i.e., up to 3.0
in. in diameter. To project these values to structural
sizes of concern, we devised an extrapolation
technique based on the published data and established information for the damping of water sloshing
in pools. This gave the damping values for the
structures of concern shown in Table 7. We will
emphasize that these values apply only to situations
in which these structures can be considered single
64

isolated members. The damping values for multiple

members can be very different.
For multiple rigid members under seismic and
normal steam-relief excitations, the concept of added
mass and added damping seems also to apply;
although the experimental confirmation is far less
extensive than for single isolated members. If we
accept the concept's validity then the added mass
effect can be calculated using potential theory.
Analytical description of the added mass effect is
more complex than for a single isolated member; it
involves "self-added" and "added" mass coefficients.
The first characterizes the force on a member from its
own motion with other members stationary, while
the second characterizes the force on a stationary
member from the motion of other members. Some
published values for these coefficients are available
for certain simple multiple member arrangements
(Refs. 10, 18,19,26,27,30,31,32,34,35,36, and 42).
An extensive compilation covering all configurations
of interest would be a major analytical undertaking
because the coefficient values are influenced by the
member arrangement, the space between members,
and the geometry of the individual members.
Reference 26 has perhaps the most extensive compilation of values.

An approximation was suggested in Ref. 35 to help

simplify the analysis cf large arrays: i.e., only the
hydrodynamic coupling between a member and its
immediate neighbors needs to be considered.
Coupling with members farther away may be
ignored. This allows an array to be analyzed in
subparts. The approximation was founded on
theoretical results for the central member of
hexagonal arrays; therefore, we are not certain how
valid it would be for peripheral members. In
particular, the corner members of an ;irr;iy receive
the highest loads, and this technique may not apply.
No experimental confirmation of the approximation
was given.
An approximation that is even simpler was
suggested in Ref. 10. The total interaction is
considered to be the sum of interactions between two
adjacent members. Therefore, the members of ihe
array are analyzed two at a time and the results
superimposed. The approximation was shown to be
accurate to within 2 to 25% for a 4 x 4 array
compared to a rigorous analysis. The accuracy
varied, depending on the member of the array.
Corner members can be analyzed.
As the gap between members of an array is
increased beyond a certain point, the members
respond as if single and isolated. For circular
cylindrical members, when the gap reaches 1.5 times
the member diameter, the members can be treated as
if single and isolated. This 1.5 value was applicable to
virtually all multiple member arrangements we
found for circular cylindrical members.
In the case of coaxial rigid cylinders, the potential
theory solution can be expressed very conveniently
for design applications. The inertial forces are given
in terms of the mass of the fluid displaced by the inner
cylinder and the mass of fluid filling the interior of
the outer cylinder in the absence of the inner cylinder.
In a more sophisticated analysis, an incompressible
viscous fluid theory was used instead of the potential
theory. The results generally agreed better with
experiments than did the potential theory. This
indicates that viscosity effects may be important for
coaxial cylinders; however, the analytical expressions are more complicated than those for potential
theory. We would like to see further confirmation
before recommending the more complex theory over
the easy-to-apply potential theory.
Coaxial flexible cylinders probably provide a
more accurate model for the reactor internals than
do coaxial rigid cylinders. Analytical treatments
generally involve a compressible invicid fluid theory
and are fairly complex. Much needs to be explored
for this case before conclusions can be drawn

regarding design oriented methods. Interest in this

area is currently high.
Damping for multiple members is presently a
broad, imprecise topic, mainly because of its
dependence on member arrangement, gap size
between members, member geometry, and whether
the member motions arc in-phasc or out-of-phase. If
the gap size is not less than 0.4 times the member size,
the damping is approximately that of a single
isolated member. Thisconvcnientsimplification may
not always apply in practice, but when it does, it

eliminates the dependence of damping on member

arrangement and whether the members are moving
in-phase or out-of-phase. It holds for coaxial
cylinders as well as for arrays. Unfortunately, it is
established for very small specimens (0.5-in. diameter) rather than for structural sizes of concern.
However, we believe we can assume the 0.4 factor
also applies to larger structures.
For gap sizes less than 0.4 times the member size,
the damping increases rapidly with decreasing gap
size when the members are moving out-of-phase with
each other. On the other hand, if the members are
moving in unison, the damping is very low, i.e., more
in the range of damping values for single isolated
members. Damping values for small gaps are
available for small specimens, i.e., 0.5-in. diameter.
At present, there is no established way to extrapolate
these values to the structural sizes of concern. The
analytical treatment presented in Ref. 27 fairly
successfully predicted the damping for 0.5-in.diameter specimens. The same procedure could be
applied to large structures; however, the success of
doing this would need to be explored.

9.2 Spent-Fuel Storage Racks

The fuel elements in a fuel bundle constitute an
array of multiple members. The cans in a spent-fuel
storage rack also form an array. The racks in a spentfuel storage pool form yet another array. The added
mass and added damping concept is applicable to
these arrays under seismic excitation. The discussions we gave on this concept for single isolated
members and arrays are directly applicable to these
structures.
The racks are generally quite stiff so, more than
likely, the entire rack will move in unison. If the rack
is isolated, the added damping would be quite low,
on the order of the values for fuel bundles shown in
Table 7. Usually, however, the rack is next to other
racks, or next to a wall. In this case, out-of-phase

65

motions between a rack and a neighboring structure

could measurably increase the damping. Unfortunately, because of a lack of information, we are not in
a position to recommend a damping value under this
condition. On the other hand, in most cases, the
racks arc firmly anchored to the pool structure and to
each other; consequently, out-of-phasc motions
would generally not occur. For this situation, the
damping would be very low, on the order of the
values for fuel bundles shown in Table 7.
Damping from component impnr* *nrt the
anchoring el feet of water need to be addressed
because they may be brought up as arguments
against the use of such low damping values for racks.
Component impact may occur between the fuel
bundles and the cans, and, according to Refs. 41 and
43, component impact can contribute measurably to
damping. However, without an indication of how
much it contributes and how much it can vary from
one rack design to another, we are not in a position to
recommend a value for it. Because of the surrounding
water, racks in a pool will tend to translate with the
pool under seismic excitation. This is an anchoring
effect, and it is a manifestation of inertial forces
rather than damping. If the added mass effect is
analyzed using the "self-added" and "added" mass
coefficients described in Section 6.2 of this report,
the anchoring effect of the water would be properly
taken into account.

9.3

Main Steam-Relief Valve Line

The added mass and added damping concept can

be applied to the main steam-relief valve line under
seismic and normal steam-relief excitations. The line
is submerged near the wall of the pressure
suppression pool. If the gap between the line and
pool wall is greater than 1.5 times thediameter of the
line, the added mass can be evaluated as if the line is
single and isolated. Otherwise, the presence of the
wall needs to be taken into account, and the curve in
Fig. 44 can provide the added mass coefficient for
motions of the line in any direction.
If the gap is greater than 0.4 times the diameter of
the line, the added damping can be assumed to be
that of a single isolated member, and the values given
in Table 7 apply. The damping will be greater if the
gap is smaller. Unfortunately, without sufficient
published information, we are not in a position to
recommend damping values for this latter case.

66

im

1r-q

1rr

2.2

1.0

0.01

' ''

'

0.1

i i i

1.0

10

G/R
KJg. 44. Hydrodynamic mass coefficient for a cylinder vibrating
near a wall.2'

9.4

Internals of the Reactor Vessel

Seismic excitation and blowdown accidents are

two important concerns for the internals of the
reactor vessel. The blowdown accident is at least an
order of magnitude more complex than the other
phenomena addressed in this project, and the
analysis techniques required are significantly more
complex and sophisticated than those required for
the others. From the very beginning of the project, it
was realized that an investigation of the accident
problem would very likely be beyond the financial
and time limitations of the project. However, NRC
and we agreed to include it in the project to see what
we could find out about it, and. if our findings were
inconclusive, that would be totally acceptable to
NRC. As it turns out, our findings are generally
inconclusive.
It became increasingly clear to us as we neared the
end of the project that, indeed, a meaningful
investigation of the accident problem is beyond the
financial and time resources of the project. Moreover, it was beyond the guidelines of the project in
that the work required to address the accident
problem is research rather than an evaluation of
methods for design calculations. The project's guidelines55 were, in essence, to investigate analytical
techniques in current use or having a potential use for
practical design calculations. Because we tailored
our efforts to this guideline, we subsequently did not
come across references which dealt in depth with the
accident problem. Our findings are generally more
applicable to seismic and normal operation excitations.

Analytical models commonly used for the internals of the reactor vessel include:
(1) two coaxial rigid cylinders.
(2) two coaxial cylinders with the inner cylinder
flexible and the outer one rigid.
(3) three coaxial cylinders with the two inner
cylinders flexible and the outer one rigid.
Model (I) is generally used to approximate the
first beam mode of the core barrel; models (2) and (3)
are to approximate shell-bending modes. Model (3)
models reactors with a thermal shield.

ratios; i.e., for D/d less than 1.8, where Dand d arc
the outer and inner diameters, respectively, of
coaxial rigid cylinders. We are not implying that Fig.
41, which is for 0.5-in.-diameter specimens, necessarily applies to sizes of a core barrel; however, the
core barrel values measured are at least not
contradictory to those in Fig. 41. Other damping
values given in Ref. 41 for reactors includes both the
effects of fluid viscosity and component impact.
These ranged from 8.8 to 12%, and no separation
with respect to the two contributions was made. We

The added mass and added damping concept can

believe the damping from component impact i;>

be applied to model (1). The potential theory expressions for the added mass response are quite
convenient for design applications and are characterized by fluid mass quantities that are simple to
determine, as explained in Section 6.3 of this report.
This technique is suitable for analysis of the first
mode beam bending deformation of the core barrel
under seismic or normal operation excitations. In a
blowdown accident the hydrodynamic effects are
most likely too severe (or this treatment; however,
the technique may provide a very rough estimate of
the response. Because no experimental confirmation
is available, we cannot make a more definite
statement concerning blowdown accidents.
A more sophisticated analysis of model (1) using
an incompressible viscous fluid theory is also
available as explained in Section 6.3 of this report.
The results seem to compare better with experimental results than did the added mass concept;
however, the expressions are more complex. This is
another technique suitable for analyzing the first
mode beam-bending deformation of the reactor core
barrel under seismic and normal-operation excitations. The same comments relating to the accident
condition given in the.preceding paragraph also
apply here.
Models (2) and (3) are generally examined with a
compressible invicid theory as discussed in Section
6.4 of this report. These models are appropriate for
shell bending modes of the core barrel and thermal
shield under seismic or normal operation excitations.
Again, the technique might be feasible for an
approximate analysis of the shell-bending deformation under an accident, provided the fluid properties are properly adjusted.
Because the annular gap between the core barrel
and the reactor vessel is small, less than 0.4 times the
diameter of the core barrel, damping from water
viscosity can be expected to be important. Total
(structural plus added) damping measured on actual
reactors revealed 2 to 5% for core barrel beam
modes, and 1 to 2% for shell modes.41 These values
are comparable with values in Fig. 41 for small D/d

bound to vary from one reactor design to another;

therefore, no reliable value of damping can be
assigned to it. In addition, component impact must
be, at least in most cases, an undesirable phenomenon that is to be avoided if possible. Therefore, if the
sought-after condition is such that component impact
is virtually absent, the fluid viscosity would be the
dominating cause of added damping. Consequently,
we would recommend using a total (structural plus
added) damping value of 2 to 5% for beam modes,
and 1 to 2% for shell modes.

9.5 Methods for Current Design Analysis

of Fuel Racks*
As mentioned in Section 4 of this report, the
methods in current use for design analysis shown in
Table 5 are based on engineering judgment together
with analytical and/or experimental information
available at the time. We will provide an assessment
of validity of each method as compared with our
suggested method based on the material presented
earlier in this report. Because of the complexity of
multiple structure-water interaction, an accurate
assessment needs rigorous analyses of the types
described in Section 6 of this report. However,
because the methods listed in Table 5 are numerous,
this is an effort beyond the scope of this project.
Consequently, our assessment is based on observations only.
The methods described in Table 5 are formulated
for low-amplitude dynamic phenomenon, such as
seismic excitation. Our assessment focuses on fuel
racks, for which many of the methods are
formulated; however, our conclusions are not
necessarily limited to fuel racks. Three types of racks
of particular interest to NRC are:
Type 1: A regular array of 9 x 9 in. cans spaced
4 in. apart.
Type 2: A regular array of 6 x 6 in. cans spaced
I in. apart.
* These methods are described in Table S.
67

Type 3: A regular array of cans with no space

between cans.
The effective mass of a submerged can is the sum of
the mass of the can, the mass of the contained fuel
rods, the mass of the water contained within the can,
and the added mass from interaction with the
surrounding water. The terms added mass and added
damping, as commonly used, pertain only to the
interaction with the surrounding water.
We will begin our discussion by first describing
I I I't rprommenHatiort"; for fuel racks. This will be
lollowed by oui assessment of the incthuds in Tabic
5.
9.5.1

situations arc shown in Table 28. Our recommendation is to use the coefficient values of the second
situation.
For the three rack types, we are, in essence, saying
the water between the cans translates directly with
the cans. Accordingly, as far as the water-structure
interaction of an entire rack module is concerned, the
module interacts essentially as a solid structure.
Therefore, in addition to the added mass for each
can, the added mass effect on the module should be
accounted for. For the three rack types of interest it
should be evaluated assuming the module is n solid
structure. To our knowledge, the space between rack
modules is generally small compared to the module
dimensions. Assuming the rack modules translate in
unison, we would recommend adding the mass of the
water between modules to the mass of the modules.
If the cans do not translate in unison, the situation
becomes significantly more complex. The analytical
method we recommended in Sections 9.1 and 9.2 of
this report should then be used. The modules,
however, can still be taken as solid structures for the
three rack types of interest.
The added damping for fuel racks is a complex
issue, and there is insufficient published technical
information on which to base a sound recommendation. Further comments are given in Section 9.2 of
this report. A major difficulty is the use of the simple
added damping concept to approximate a phenomenon that is measurably more complex in the case of
multiple members. Unfortunately, no better alternatives were found for a design-oriented approach.
Therefore, we currently suggest using the added
damping values given in Table 7; i.e., 0.6% for rack
type 1, and 0.4% for rack type 2, and 0% for rack
type 3. based on the type 3 rack module responding
as a unit. We recommend that further studies on
damping be carried out, particularly experimental
studies.

LLL Recommendations for Fuel Racks

If the fuel racks arc arranged so that the predominate modes of vibration consist of the cans
translating in unison, we recommend using the coefficients in Table 27 to evaluate the added mass per
can. The added mass per can is the coefficient times
the mass of the water displaced by the exterior
volume of the can.
The basis for our recommendations, in the absence
of a rigorous analysis, is that we believe the added
mass per can in an array should be the smaller of
either the added mass for the can if single and
isolated or the mass of the water actually surrounding the can in the array. For the three rack types of
interest, the coefficient values for these two

Table 27. Added mass coefficients for evaluating

added mass per can.
Rack type

Added mass coefficient

1.086
0.36

Table 28. The coefficient values for three rack types for (1) the added mass of a single isolated can and
(2) the mass of water surrounding the can.
Rack
type

Potential theory
for situation 1
1.186
1.186
1.186

13 X 13 9X9
= 1.086
9X9
7 X 7 6X6
For 6 X 6 in. cans spaced 1 in. apart,
= 0.36
6X6

* For 9 X 9 in. cans spaced 4 in. apart,

68

Actual surrounding water

Tor situation 2*
1.086

0.36
0

9.5.2 Methods No. 1 and No. 2

For single isolated members, potential theory will
provide valid added mass values. Using the expressions tabulated in Tabic 6 is the same as applying
potential theory.
Applying potential theory to evaluate the added
mass for multiple members is likewise valid. However,
the solution procedure is usually difficult, and approximate solutions are normally sought. The accuracy of the approximate analytical model may be of
primary concern. ijeuuis on liie approximations
used, if any, and on the damping values used were not
given by the firms using Methods No. I and No. 2.
9.5.3

Our objections to using a total damping of two

times the structural damping was discussed in
Section 4 of this report.
9.5.4.

Method No. 4

Method No. 4 does not appear to be extracted

from the references (Rcfs. I, 3, 4, 7, and others)
mentioned by the user. For single isolated square
members, the added mass is approximately one-third
that given by potential theory. For the rack types of
intercut the iiddccl m;:" per enn i* "i v cr ^ v *1'*
coellicients values in I'able o'J, in comparison with
LLL's recommendations. It appears Method No. 4
will significantly underestimate the LLL recommendation for added mass for rack type I.
Based on the information available on added
damping the use of zero added damping is certainly
conservative.

Method No.3

For single isolated members, Method No. 3 gives

the same results as the procedure recommended by
Newmark and Rosenblueth (N&R).5 Therefore, as
discussed in Sections 5.1 and 5.2 of this report, the
added mass value would be two-thirds that given by
potential theory for a square member. We would
recommend using potential theory over Method No.
3 for greater accuracy, for essentially the same level
of analytical complexity in the case of single isolated
members.
For multiple members, Method No. 3 does not
appear to be extracted from the references (Refs. 1,3,
and others) mentioned by the user. For the three rack
types of interest the resulting coefficient values are
shown in Table 29, in comparison with LLL's
recommendations. It appears Method No. 3 will
significantly underestimate the LLL recommendation for added mass for rack types 1 and 2.

9.5.5.

Method No. 5

Potential theory will provide valid added mass

values for single isolated members.
For multiple members, the user is addressing more
general motions than unison translation. Fritz's
method7 provides an approximation for a generic
member surrounded by adjacent members. The
procedure by which the user applied Fritz's method
to an array was not totally clear to us. However, for
unison motion, we speculated that recognition was
probably given to the fact that a generic member
plays two roles: as the central member and as part of
the surrounding square cylinder for its adjacent
members. Therefore, superimposing the forces given

Table 29. Coefficient values for three rack types comparing Method No. 3 of Table 5 with the LLL
recommendation.
Rack
type

Table 30.
Rack
type

Method No. 3
coefficients

I.I.I.-recommended
coefficients

0.44

1.086

0.17

0.36

Coefficients for three rack types comparing Method No. 4 with the LLL recommendation.
Method No. 4
coefficients

I.LI.-recommended
coefficients

0.4

1.086

0.36

0.36

69

Table 31.

Coefficients for three rack types comparing Method No. 5 with the LLL recommendation.
Method No. 5
coefficients*

Hack
type

XL-recommended
coefficients

2.S68

1.086

0.778

0.36
0

* For 9 X 9 in. cans spaced 4 in. apart,

For 6 X 6 in. cans spaced 1 in. apart,

17X 1 7 9 X 9

9X9
8 X 8 - 6 X 6

6X6

= 2.568

= 0.778

in Fritz7 for the two roles played by a generic member

lead us to conclude that the added mass for a given
member is the mass of the water between the member
and its surrounding square cylinder. This would
result in the added mass coefficients for the rack
types of interest shown in Table 31, in comparison
with LLL's recommendations. Consequently, according to our speculation of how the user applied
Fritz's method, Method No. 5 overestimates the LLL
recommendation for added mass for rack types 1
and 2.
The added damping ranged from 0 to 3% as
compared with our previously recommended value.
Because of insufficient information on their use of
added damping, we are in a poor position to assess
the validity of these values.
9.5.6.

Method No.6

An added mass equal to the mass of water

displaced by the exterior volume of a square can will
be 0.843 times that given by potential theory, for
single isolated square cans. We prefer using potential
theory, for essentially the same analytical complexity.
For the three rack types of interest, the procedure
for evaluating the natural frequency corresponds to
LLL's recommendation. However, the procedure for
evaluating inertial loads essentially ignores the added
mass. Consequently, we consider Method No. 6
inappropriate (in comparison with LLL recommendations) for analyzing inertial effects for rack
types I and 2.

70

An added damping of 2% does not correspond to

the value we previously suggested.
9.5.7.

Method No. 7

For single isolated members, the comments made

for Method No. 6 also apply to Method No. 7. For
multiple members translating in unison, Method No.
7 corresponds almost to LLL's recommendation.
The only difference is LLL recommends using
potential theory rather than the mass of the displaced
water to evaluate the added mass for single isolated
members. For the three rack types of interest,
Method No. 7 and LLL's recommendations give the
same values for the added mass.
We discourage the use of 2 to 2-Vi times the
structural damping as the total damping. An added
damping of 2% does not correspond to the value we
previously suggested.
9.5.8.

Method No. 8

We consider Method No. 8 inappropriate (in

comparison with LLL recommendations) for evaluating the added mass effect for rack types I and 2. We
disagree with the use of two times the structural
damping as the total damping for the reason
discussed in Section 4 of this report.
9.5.9

Method No. 9

The user is apparently involved with only rack

type 3, and Method No. 9 is appropriate for this
configuration. The use of zero, added damping could
be appropriate and is certainly conservative.

10.

ACKNOWLEDGMENTS

The author expresses his appreciation to G. A.

Broadman, Leader, C. E. Walter, Deputy Leader,
and F. J. Tokarz, Associate Leader, of the Nuclear
Test Engineering Division, Mechanical Engineering
Department, for their encouragement and support of
this project. The project was initiated, sponsored.

and funded by the U.S. Nuclear Regulatory

Commission. The attention and advice given by Dr.
C. H. Hofmayer and Dr. K. S. Herring of the
Division of Operating Reactors has been most
helpful and appreciated,

11. REFERENCES
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
16.
17.
i 8.
19.
20.
21.
22.

R. W. Clough, "Effects of Earthquakes on Underwater Structures," Proc. of 2nd World Conference on

Earthquake Engineering. (Tokyo, I960).
K. T. Patton, Tables of Hydrodynamic Mass Factors for Translational Motion, ASME Paper
No. 65-WA/UNT-2.
A. R. Chandrasekaran, S. S. Saini, and M. M. Malhotra, "Virtual Mass of Submerged Structures,"
Journal of the Hydraulics Div.. Proc. of the ASCE. (May, 1972).
T. E. Stelson and F. T. Mavis, Virtual Mass and Acceleration In Fluids. ASCE Trans. Paper No. 2870,
Vol. 122, 1957.
N. M. Newmark and E. Rosenblueth, Fundamentals of Earthquake Engineering, (Prentice-Hall Inc.,
Englewood Cliffs, N. J. 1971), Chapter 6.
G. H. Keulegan, and L. H. Carpenter, "Forces on Cylinders and Plates in an Oscillating Fluid," Journal of
Research of the National Bureau of Standards, 6 (5), (May, 1958).
R.J. Fritz, "The effects of Liquids on the Dynamic Motions of Immersed Solids," Journal of Engineering
for Industry, Trans. ASME (Feb., 1972).
G. B. Wallis, One-Dimensional Two-Phase Flow (McGraw-Hill Book Co., New York, 1969), Chapter 8.
L. S. Jacobsen, "Impulsive Hydrodynamics of Fluid Inside a Cylindrical Tank and of Fluid Surrounding
a Cylindrical Pin," Bull, of the Seismological Society of America. 39 (1949).
T. Yamamoto and J. H. Nath, "Hydrcdynamic Forces on Groups of Cylinders," Offshore Technology
Conference, 1976 Proc.. Vol. I. OTC 2449.
H. B. Amey, Jr., and G. Pomonik, "Added Mass & Damping of Submerged Bodies Oscillating Near the
Surface," Offshore Technology Conference. 1976 Proc., Vol. 1, OTC 1557.
C. J. Garrison, and R. B. Bcrklitc,"Hydrodynamic Loads Induced by Earthquakes," Offshore Technology
Conference, 1976 Proc.. Vol. I. OTC 1554.
T. Sarpkaya, "Separated Flow About Lifting Bodies and Impulsive Flow About Cylinders," AIAA
Journal, 4 (3), (March, 1966).
A. Selby and R. T. Severn, "An Experimental Assessment of the Added Mass of Some Plates Vibrating
In Water," Earthquake Engineering and Structural Dynamics, 1 (1972).
H. Schlichting, Boundary Layer Theory (McGraw-Hill Book Co., New York, 1968).
J. H. Shames, Mechanics of Fluids (McGraw-Hill Book Co., New York, 1972).
A. R. Chandrasekaran and S. S. Saini, "Vibration of Submerged Structures," Irrigation and Power (July,
1971).
T. Yamamoto, "Hydrodynamic Forces on Multiple Circular Cylinders," Journal of the Hydraulics Division, Proc. of the ASCE (Sept., 1976).
C. Dalton and R. A. Helfinstine, "Potential Flow Past A Group of Circular Cylinders," Journal of Basic
Engineering, Trans. ASME (Dec, 1971).
R. A. Skop, S. E. Ramberg, and K. M. Ferer, Add Mass and Damping Forces on Circular Cylinders,
ASME Paper 76-Pet-3, 1976.
S. Levy, J. P. D. Wilkinson, "Calculation of Added Water Mass Effect for Reactor System Components,"
Trans, of the 3rd Int. Conf. on Structural Mechanics in Reactor Tech. (Sept., 1975).
J. G. Alibaud, (Abstract Only) "Virtual Mass Effect of Water on the Internals of Pressurized Water
Reactors, Theory and Experimental Results," Trans, of the 3rd Int. Conf. on Structural Mechanics In
Reactor Tech. (Sept., 1975).
71

23.

24.

25.

26.
II.
28.
29.
30.
31.
32.
33.
34.
35.
36.
37.
38.
39.

40.

41.
42.

43.

44.

45.
46.

72

R. Assedo, M. Dubourg, M. Livolant, A. Epstein, "Vibration Behavior of FWR Reactor Internals

Model Experiments and Analysis," Trans, of the 3rd Int. Conf. on Structural Mechanics in ReactorTech.
Sept. 1975.
C. Carmignani, F. Manfrcdi, S. Rcale, P. Cewetti, A. Fcdevico, "Influence of the Assembly Configurations on the Flow Induced Vibrations of BWR Fuel Elements," Trans, of the 3rd Intern. Conf. on
Structural Mechanics In Reactor Technology (Sept., 1975).
D. T. Ramani, "Stress and Deflection Analysis of Reactor Internals Due to Seismic & Maximum Hypothetical LOCA Loading Conditions," Trans of the 3rd Inter. Conf. On Structural Mechanics In Reactor
Technology (Sept., 1975).
S. S. Chen and H. Chung, Design (iuidefor Calculating Hydrodynamic Muss. Purl I: Circular Cylindrical
Strut-turn*. An>onne National Laboratory ANL-CT-76-45 (June, 1976).
S. S. Chen and M. W. Wamb.sgans.s, and J. A. Scndo/cjc/yk, "Added Mass ;md Damping r>t a Vibrating
Rod In Confined Viscous Fluids," J. of Applied Mechanics, 98 (2) (June, 1976).
M. Ohkusu, "Wave Action on Groups of Vertical Circular Cylinders," Proc. of the Annual Spring
Conf. of Japan Sue. of Naval Architects (May, 1972).
T. Sarpkaya, "Forces on Cylinders and Spheres In A Sinusoidally Oscillating Fluid," Trans. ASME. J. of
Applied Mech. (March, 1975)
P. M. Morctti, and R. L. Lowery, "Hydrodynamic Inertia Coefficients for a Tube Surrounded by Rigid
Tubes," Trans ASME J. of Pressure Vessel Technology (August, 1976).
Y. S. Shin and M. W. Wambsganss, "Flow-Induced Vibration in LMFBR Steam Generators: AStatc-ofthe-Art Review," Nuclear Engineering and Design, 40 (1977), pp. 235-284.
S. S. Chen, "Vibration of Nuclear Fuel Bundles," Nuclear Engineering and Design. 35(1975), pp. 399-422.
L. W. Carpenter, "On the Motion of Two Cylinders in An Ideal Fluid," Journal of Research of the
National Bureau of Standards. 61 (2) (August, 1958), Research Paper 2889, pp. 83-87.
H. Chung and S. S. Chen, Vibration of a Group of Circular Cylinders in a Confined Fluid, ASME Paper
No. 77-APM-I6 (1977), pp. 1-5.
S. S.Chen, Dynamics of Heat Exchanger Tube Banks. ASME Paper No. 76-WA/FE-28( 1976), pp. 1-7.
S. S. Chen, "Vibrations of a Group of Circular Cylindrical Structures in A Liquid," Trans, of the 3rd International Conf. on Structural Mechanics in Reactor Technology, Vol. I Part D(September, 1975), pp. 1-11.
M. Dubourg, R. Assedo, and C. Cauquelin, "Model Experimentation and Analysis of Flow-Induced
Vibrations of PWR Internals," Nuclear Engineering and Design, 27 (1974), pp. 315-333.
M. K. Au-Yung, "Response of Reactor Internals to Fluctuating Pressure Forces," Nuclear Engineering
and Design. 35 (1975), pp. 361-375.
D. T. Ramani, "Transient Dynamic Response and Stability Analysis of Reactor Core Support Barrel
Due to LOCA Pressure Pulse Loading," Trans. 4th Int. Conf. on Structural Mechanics in Reactor
Technology (San Francisco, California, August, 1977) (Abstract available only).
R. Longo, F. W. Martsen, P. A. Perrotti, E. C. Bair, "Seismic Analysis of Spent Fuel Racks," Trans. 4th
Int. Conf. on Structural Mechanics in Reactor Technology (San Francisco, California, August, 1977)
(Abstract available only).
N. R. Singleton, and G. J. Bohm, "Damping of Reactor Internals," Trans 4th Int. Conf. on Structural
Mechanics in Reactor Technology (San Francisco, California, August, 1977).
R. W. Wu, L. K. Liu, S. Levy, "Dynamic Analysis of Multibody System Immersed in a Fluid Medium,"
Trans. 4th Int. Conf. on-Structural Mechanics in Reactor Technology (San Francisco, California,
August, 1977).
Y. S. Shin, J. A. Jendrzejczyk, and M. W. Wambsganss, "The Effect of Tube-Support Interaction on the
Dynamic Response of Heat Exchanger Tubes," Trans. 4th Int. Conf. on Structural Mechanics in Reactor
Technology (San Francisco, California, August 1977).
J. J. Dubois, A. L. deRouvray, "Improved Coupled Euler-Lagrange Finite Element Analysis of the
Fluid-Structure Dynamic Interaction Problem," Trans. 4th Int. Conf. on Structural Mechanics in
Reactor Technology (San Francisco, California, August, 1977).
R. J. Fritz and E. Kiss, The Vibration Response of a Cantilevered Cylinder Surrounded by An Annular
Fluid, KAPL-M-6539, General Electric Co., Schenectady, N.Y. (February, 1966).
M. K. Au-Yang, "Free Vibration of Fluid-Coupled Coaxial Cylindrical Shells of Different Length," J. Of
Appl. Mech. 98 (3), (Sept., 1976).

47.

48.
49.
50.
51.

T. M. Mulcahy, P. Turula. H. Chungand.l. A..Icndrzcjc/yk, "Analytical and F.xpcrimcntal Study of Two

Concentric Cylinders Coupled by a Fluid Gap," Trans. 3rd Int. Conf. on Structural Mechanics in
Reactor Technology (London. Sept.. 1975).
S. S. Chen, J. A. Jendrzejczyk, and M. VV. Wambsganss, Experiments on Fluid Elastic libraiions of
Tube Arrays, ANL-CT-77-16, Argonnc National Laboratory (April, 1977).
T. T. Yeh and S. S. Chen. The Effect of Fluid I iscosity on Coupled Tube I Fluid I ibrations, A N L-C 1-7724, Argonne National Laboratory (April. 1977).
U.S. Nuclear Regulatory Commission. Regulatory Guide 1.60. Design Response Spectra for Seismic
Design of Nuclear Power Plana (Dec, 1973).
H. N. Abramson, The Dynamic Behavior of Liquids in Moving Containers. National Aeronautics and
Space Administration. Rept. NASA SP-106 (1966).

52.

53.

54.
55.

D . D . N a n a , a n d r . i . D o d g e . " D e s i g n S u p p o i l iviouciuig oi L i q u i d S l o s h in S l o i a g c t a n k s Sui>|ccl U>

Seismic Excitation," presented at Second A SCE Specially Conf. on Structural Design of \uclear Plant
Facilities. (New Orleans. LA., Dec, 1975).
D. G. Stephens, H. W. Leonard, and T. W. Perry, Jr., Investigation of the Damping of Liquids in RightCircular Cylindrical Tanks, Including the Effects of a Time-Variant Liquid Depth, NASA IN D-1367
(July, 1962), Wash. D.C.
D. D. Kana, of the Southwest Research Institute. San Antonio. Texas, Personal communications. August
18, 1977, in San Francisco, California.
Project Proposal, Effective Mass and Damping of Submerged Structure. Contract No. B&R 20 19 04 02
FIN AO203, August 1976.

KJM

G P O 79S-072/&86

73