Calhoun: The NPS Institutional Archive

DSpace Repository

Faculty and Researchers Faculty and Researchers Collection

1979-02

Wave Impact Loads on Cylinders

Sarpkaya, Turgut
Society of Petroleum Engineers of AIME

http://hdl.handle.net/10945/49131

Downloaded from NPS Archive: Calhoun

 Wave Itppact Loads on Cylinders
NAVAL POSTGRADUATE SCHOOL
TURGUT SARPKA Y A
MONTEREY, CA

ABSTRACT Kaplan and Silbert 2 developed a solution for the

forces acting on a cylinder from the instant of
The evolution of forces acting on horizontal impact to full immersion. Dalton and Nash 3
cylinders subjected to impact by a sinusoidally conducted slamming experiments with a 0.5-in.
oscillating free surface was investigated both (12.7-mm) diameter cylinder and small amplitude
theoretically and experimentally. The experiments waves created in a laboratory tank. Their data
were conducted in a large V-shaped tunnel, with exhibited large scatter and showed no particular
cylinders 3 to 8 in. (76 to 203 mm) in diameter. correlation with either the predictions of the
The results are expressed in terms of three force hydrodynamic theory or identifiable wave parameters.
coefficients: (1) the general slamming coefficient Miller 4 presented the results of a series of
that expresses the normalized force acting on the wave-tank experiments to establish the magnitude
cylinder at any time after the impact, (2) the of the wave-force slamming coefficient for a
normalized impact force at the initial instants of horizontal circular cylinder. He found an average
slamming, and (3) the maximum drag coefficient slamming coefficient of 3.6 for those trials 10
that occurs when the cylinder is immersed in which slamming was dominant.
water about 1. 8 diameters. The slamming-force Evaluating slamming effects with wavy flows is
coefficient was found to equal 3.2. Also, the force extremely difficult partly because of the limited
experienced by the cylinder cannot be considered range of wave amplitudes that can be achieved and
independently of the dynamic response of that· partly because of the difficulty of measuring the
cylinder. In fact, the slamming-force coefficient fluid velocities at the instant of impact.
may be amplified to a value as high as 6.3 through Faltinsen et al. S investigated the load acting on
the dynamic response of the cylinder and its rigid horizontal circular cylinders (with end plates
supports. and length-to-diameter ratios of about 1) that were
forced with constant velocity through an initially
INTRODUCTION calm free surface. They founq. that the slamming
Infonnation about the forces acting on bluff coefficient ranged from 4.1 to 6.4. They also
bodies subjected to wave slamming is of significant conducted experiments with flexible horizontal
importance in ocean engineering and naval cylinders and found that the analytically predicted
architecture. The design of structures that must values were always lower (50 to 90%) than those
survive in a wave environment depends on a found experimen tall y.
knowledge of the forces that occur at impact, as This investigation was undertaken (1) to examine
well as on the dynamic response of the system. the existing theoretical models for detennining
Two typical examples include the structural wave slam forces on circular cylinders; (2) to
members of offshore drilling platfonns at the splash furnish data, obtained under controlled laboratory
zone and the often encountered slamming of ships. conditions, about forces acting on circular cylinders
The general problem of hydrodynamic impact has subjected to impact with a sinusoidally oscillating
been studied extensively, motivated in part by its water surface; (3) to determine the relative
importance in ordnance and missile technology. 1 importance of the inertia- and drag-dominated
Extensive mathematical models have been developed forces during fluid impact; and (4) to correlate
for cases of simple geometry, such as spheres and these data for identifiable wave parameters such
wedges. These models have been well supported by ~s the Froude number (N Fr ); the Keulegan-Carpenter
experiment. Unfortunately, the special case of number (N K ); and the Reynolds number (N Re ).
wave impact has not been studied extensively. This investigation does not deal with the
relatively more complex impact situations arising
Original manuscript received in Society of Petroleum Engineers
office March 13, 1977. Paper accepted for publication Sept. 14,
from the slamming of random ocean waves. on the·
1978. Revised manuscript received Oct. 5, 1978. Paper (SPE members of offshore structures. Such practical
7216, OTC 3065) first presented at the Tenth Annual Offshore
Technology Conference, held in Houston, May 8-11, 1978. problems require not only a clear understanding of
This paper will be included in the 1979 Transactions volume. the origins and magnitude of the impact forces
0037-9999/79/0002-7216$00.25 through idealized experiments, but also the use of
@ 1979 Society of Petroleum Engineers of AIME statistical design methods and some knowledge of
FEBRUARY 1979
the probabilities associated with varIOUS force in which r represents the radius of the cylinder,
and response coefficients. and e is defined as shown in Fig. 1.
The motion of the free surface, "I, is related to
THEORETICAL ANALYSIS the maximum ampli tude by

The traditional approach used to design offshore n = A sin 27ftjT, . . . . . . . . (3)

structures involves the classical Morison equation
for determining the forces caused by wave motion. where A and T represent the amplitude and period
Theoretical and practical aspects of the determina- of the free-surface oscillations. Eq. 1 also can be
tion of wave forces on offshore structures .were written in the form of a slamming coefficient, Cs =
2F/pDU1 m 2 , as
described in detail by Sarpkaya and Isaacson
recently. 6
Wave impact generally is defined as the early cs = A;. ~ - (m· + A ) r . 27f t
stages of the penetration of a solid body into a ~ U2 i A slnlf
wave surface .. At the instant of contact, the fluid m
near the body undergoes large accelerations that
create large forces .. As the body becomes more dm 2 27f t ,
fully immersed, forces resulting from viscous drag + cos T (4)
and separation effects dominate. Thus, the inertia d4
and drag coefficients used in Morison's· equation where
are not constant, and the problem becomes very
difficult, even for simple geometries. A. = A.1 jr 2 m = mj7fr 2
The general case of hydrodynamic impact usually 1
is described by using incompressible potential
flow theory. For a moving body with mass, M, and Z = zjr U
m = 27fAjT
velocity, vo' impacting a quiescent surface, the
system momentum is Mvo. Neglecting nonconserva- With z = r[l cos(e/2)], Eq.2 can relate the
tive forces, the momentum of the system is added mass to the depth of immersion .. Clearly,
unchanged during penetration. However, the mass aiii/az begins with an initial value of 77 and drops
of the system increases because of the fluid set in rapidly .. The quantity gr/U~ is related to the
motion near the body. Also known as added mass, Froude number by NFr = U~/2gT. Thus, Cs =
m results in reducing the velocity .. Thus, the f(N Fr , r/ A, amjaz).
system momentum after penetration is (M + m) v = The rate of change of the normalized added mass
Mvo. The impact force at any instant is a function with z depends on e
and, hence, on the time
of m and am/at .. Therefore, the solution requires measured from the instant of impact. For example, for
knowledge of the added mass and its time derivative. the case where "10 = 0 and t = 0, ami az
= 77, and
As shown by Kaplan and Silbert, 2 the force acting C~ = 77. Consequently, for the particular case under
on a horizontal circular cylinder by a wave system consideration, C~ at the instant of impact does not·
that propagates normal to it is equal to the sum of depend on either the size of the cylinder or the
the buoyant force and the time-rate of change of flow parameters.
momentum. Thus, one has Because the cylindrical members of a structure
in the splash zone are not necessarily located at
the mean water level requires the determination of
the particular value of "10 for which there is
(1) maximum slamming force at the time of impact.
Using Eq. 4, it can be demonstrated that the
maximum impact force occurs for the case of "10 =0.
in which F represents the force acting on the
For this purpose, Eq. 4 was evaluated with the aid
cylinder; L, the length of the cylinder; p, the fluid
density; g, the gravitational acceleration; A i, the
immersed area; m, the added mass per unit length;
"I, the instantaneous height of the wave surface
above the mean water level; and z, the in~tantaneous
depth of cylinder immersion. The first and second
derivatives of "I for time are denoted by ~ and fi.
The added mass is given by Taylor7 as

~
3 t
1 2 [27f (1 - cose)
m = -pr f
2 3
(27f - e) 2
7f
+ 3(1 - cose) + sine - e]
~o
____ 1_____________L__!.!~L~
(2) FIG. 1 - DEFINITION SKETCH FOR THE
THEORETICAL ANALYSIS.

SOCIETY OF PETROLEUM ENGINEERS JOURNAL

of a computer that allowed variations In r/ A and rather on a cylinder that is supported elastically.
"'Io/A. Cs then was plotted as a function of z/D The response of such a system approaches that of
from zero to unity. Fig. 2 is a typical example of a rigid body only if its natural frequency approaches
one of these plots. The particular values used were infinity .. Additionally, the response of the system
r/ A = 0.125, and "'lolA = 0.0, 0.40, and 0.80. As to an impulsive force depends heavily on the exact
can be seen readily from Fig. 2, C s is largest at nature of the force itself as well as on the system
z/D = 0 for "'10/ A = 0, starts at a value of 1T, and natural frequency.
drops to a minimum at z/D of about 0.5. Cs rises In general, the instantaneous displacement, x(t),
again as z/ D approaches unity. Thus, Eq. 4 of a system of mass M with a spring constant k is
indicates that Cs ' arid consequently the impact given by (Ref. 8)
force, is of an implusive nature beginning with a
finite value at the instant of impact. Since viscous t
forces are neglected, one would expect the solution
to deviate from the actual situation as the cylinder
becomes more fully immersed. Where this becomes
x(t) = f F(~) g(t-~)d~ , . . . (6)
the case can be determined only experimentally.
o
However, one can assume that the solution in which ~ represents a dummy variable; pet), the
would be valid for small depths of immersion, driving force; and g(t), the response to a unit step
which is the case for the instant of impact up to excitation. Eq. 6 readily is recognized as the
z/D of 0.02 or more. With this restriction, Ai and m Duhamel superposition integral, which can be
are small, as is sin 21Tt/T. Also, cos 2 21Tt/T IS expanded as follows for get) = (l/Mw n ) sinwnt,
almost equal to 1.0. Therefore, C s reduces to
X(t) =
t
It is unrealistic to assume that the impact force
rises from zero to 1T instantaneously. Several factors,
specifically the compressibility of the air between
the cylinder and water surface, entrapped gases in the
Mw
1
n
coswnt
J
0
F(t) sin
n
t dt

water, and surface irregularities, would account t

J
for some finite rise time. Nonetheless, the rise
time can be expected to be short - i.e., in the + 1 .
~slnw
t F(t) cosw t dt
order of milliseconds. The exact nature of the rise wn n n
is an interesting question for further study. However, 0
here Cs is assumed to vary linearly during a rise
period, tr • For t > tr , Cs drops from 1T as am/az. . . . . . . . .. . . .(7)
Fig. 3 is a representation of this assumption.
in which wn is the natural frequency of the cylinder
Exactly how long a time interval tr is will be and supports.
discussed later.
The realization that· the impact force has an
implusi ve nature requires consideration that this
force does not act on a perfectly rigid body, but

4.0

z
D
o DoS 1.0

FIG. 2 - THEORETICAL VALUES OF THE I-- tr ~

SLAMMING COEFFICIENT AS A FUNCTION OF FIG. 3 - RISE TIME AND THE SLAMMING
RELATIVE SUBMERGENCE AND POSITION. COEFFICIENT AS A FUNCTION OF TIME.

FEBRUARY 1979 31
 By letting F(t) = Fo !(t) and by changing the The damping coefficient was 0.014 in both cases,
variable of integration to f3 = wnt, we obtain also corresponding to the measured values. The
rise time varied from zero to about 0.025 s. Figs.
S/W n 4 through 7 are representative plots of kx(t)/C~
k X(t)
FO
= cosS f
o
f(t) sinS dS
for the 6-in. (153-mm) cylinder for various tr •
Fig. 4 represents Eq. 11 plotted for a rise time
of 0.0001 s; Fig. 5 represents a rise time of 0.0100
s; Fig. 6 is a rise time of 0.0195 s; and Fig. 7
S /w n represents a rise time of 0.0230 s. Note that for a

+ sinS f f(t) cosS dS

very short rise time, the value of kx(t)/C~ reaches
a value of about 1. 7 at t = 0.001 s and then drops
off, with a "double peak" appearing at a rise time
o of 0.0195 s.
The interpretation of these results indicates
(8)
that, depending on the rise time, values of the
apparent slamming coefficient, C~, may lie between
in which k = w;;M, the spring constant. The term on roughly 0.5 and 1.7' of the theoretical value of 77.
the left side can be interpreted as the ratio of the Again, this applies only to a small depth of
instantaneous force sensed by the supports of the immersion corresponding to the initial moments of
cylinder to the actual force acting on the cylinder. the impact. The significance of these facts is that
For the simple case in which F (t) is a step function, values of C s determined experimentally from the
Eq. 8 reduces to measured reaction forces at the supports of the
cylinder may show wide scatter, depending on the
k X(t) = (1 - cosw n t)· . . . . (9) dynamic response of the cylinder and the rise time.
Fo Also, if the surface is not perfectly plane, rise
times may vary from experiment to experiment,
It is apparent from Eq. 9 that the instantaneous resulting in an apparent nonrepeatability. This is.
force sensed by the system can be anywhere from particularly true for conditions in the ocean
zero to two times the actual mean force. environment. Evidently, controlled laboratory experi-
If one assumes, as before, that the impact force ments help to establish the ideal value of the
is as shown in Fig. 3, then Eq. 8 must be evaluated
using Cs(t) as F(t). Additionally, damping can be
accounted for by writing get) as
2.0 ~

-sW n
t
. t . (0)
o
k x(t)/ Cs
g ( t ) =e slnw n '
tr 0.0001 sec.
where , is the damping coefficient. Thus, Eq. 8
can be rewritten by replacing the forcing function 1.51-
f\
F(t) by C s (f3) as

=
k X(t)
CO
s 1.0 ~
n
S/W n
-sS
-cosS
f
o
Cs(S) e sinS dS
0.5

f3/ Wn
-sf3
+sinS
fo Cs(S) e cosS dS
0~----4-~
I
____+-~L-~~
0.02 0.04
It
__- L
0.06
__

. . . . . . . . . . . . . (11)
Eq. 11 was solved numerically using a trapezoidal v
integration scheme. Values of wn were taken as
358 s-l and 628 s-l, which corresponded to the
measured values of wn for a 6.0-in.053-mm) and a
-0.5 ~
v
FIG. 4 - AMPLIFICATION FACTOR AS A FUNCTION
3.0-in. (76-mm) diameter cylinder, respectively. OF TIME FOR A RISE TIME OF 0.0001 s.

32 SOCIETY OF PETROLEUM ENGINEERS JOURNAL

impact coefficient and to explain the reasons for two vertical legs are 3 x 6 ft (91 x 182 cm) in
the observed scatter in the data. The determination cross-section. Previously, a butterfly valve arrange-
of the impact-force magnification factor for the ment at the top of one leg was used to start the
cylinder in the ocean environment necessarily must oscillations .. Recently, the tunnel was modified
consider the random nature of the disturbances at so that oscillations could be generated and
the wave surface, the orientation of the structural maintained indefinitely at the desired amplitude.
member for a given wave, currents, three-dimensional A 2-Hp (1.5-kW) fan was connected to the top
nature of the waves, spray, etc. In short, there are of one leg of the tunnel with a 2-ft (60-cm) diameter
no deterministic means for predicting the reaction pipe. A small butterfly valve, placed in a special
forces acting on the supports of a member resulting housing between the top of the tunnel and
from wave impact, even when the structural the 2-ft (60-cm) supply line, oscillated harmonically
characteristics of the member (damping, natural at a frequency equal to the natural frequency of the
frequency, etc.) and the ideal value of the impact oscillations in the tunnel. The oscillation of the
force are known. This is true because the rise time valve was synchronized perfectly with that of the
depends on all these nondeterministic conditions. flow using a feedback control system. A pressure
transducer (sensing the instantaneous acceleration
EXPERIMENTAL EQUIPMENT of the flow) was connected to an electronic
AND PROCEDURES speed-control unit coupled to a DC motor that
oscillated the valve plate. The circuit maintained
The equipment used to create the harmonically the period of oscillation s of the valve within
oscillating flow has been used extensively at the
Naval Postgraduate School for the past 4 years.
The apparatus is described in Ref. 9. Only the 1.5
salient features, most recent modifications, and'
k x(t)
the adaptation for this study, are described briefly
in the next section. CO
s
U-SHAPED WATER TUNNEL t 0.0195 sec.
r
The tunnel is 32.8 ft (10 m) long x 20 ft (6 m) 1.0
high with a 3 x 3 ft (91 x 91 cm) test section. The

2.0 0.5

k x(t)
CO
s
1.5 o
0.010 sec. o 0.02 0.04 0.06
FIG. 6 - AMPLIFICATION FACTOR AS A FUNCTION
OF TIME FOR A RISE TIME OF 0.0195 s.

1.0 1.5
k x(t)
CO
s
1.0 0.023 sec.
0.5

-0.5
FIG. 5 - AMPLIFICATION FACTOR AS A FUNCTION FIG. 7 - AMPLIFICATION FACTOR AS A FUNCTION
OF TIME FOR A RISE TIME OF 0.010 s. OF TIME FOR A RISE TIME OF 0.023 s.

FEBRUARY 1'179 33
± 0.001. The fluid oscillated with a period of 6.00 and
s. The amplitude of the oscillations was varied by 2
constricting or enlarging an orifice at the exit of gDn/(2U )
m
the fan. The orifice opening was calibrated vs the
. . . . . . . (13)
amplitude of flow oscillation. It was possible to
oscillate the flow at the desired amplitude as long
as desired. The former refers to the slamming coefficient as
defined by Eq. 4 and includes the contribution of
TEST CYLINDERS the dynamic response of the system .. The latter
About 3- to 8-in. (76.2- to 203-mm)-diameter represents the second maximum of the normalized
smooth and rough aluminum cylinders were used for force with the buoyancy subtracted. This maximum
measuring the impact .forces. An accelerometer was occurs after the cylinder is fully immersed. Fig .. 9
placed inside each cylinder to measure the is a typical plot of the measured value of Cs ' the
instantaneous vertical acceleration of the cylinder. theoretical value of C s from Eq. 4, and the
Caps fitted to each end prevented water leakage normalized buoyant force ..
into the cylinder .. Each cap contained a double- The third coefficient is C~, which represents the
precision ball bearing mounted flush with the face.
The force transducers were attached to the cylinder
by these bearings.
The cylinders were placed in the 3 x 6 ft (91 x
182 cm) section of one leg of. the tunnel. Since the
height of the mean water level for the bottom of the !!.
1< ,." ....... 'II'
cylinder was shown by analysis to be important, I .v ch-' ,I"-'
the water level in the tunnel was adjusted carefully , , '. I... .. "" !" 0.10 sec. ~f, ,'fS
C4--+"+HIf+'+-'-l++-H+-+-+++--1crlit"4,, t-+ .. I: co: ,."c,
..

by slowly filling and emptying the tunnel until a .,,.': 'I :::
slight ripple occurred at the water surface because ..•.. B
of contact with the cylinder. This method assured .. , ','. I,':',
that the mean water level coincided with the bottom FORCE HIH++'H1-H tft-;rilH'r.\:-tt1
1I'¥-H-t--4-+1Hit:'trli. ,riHctt\t-fr \rl:1\'r.d'".f
t f•,

surface of the cylinder. The damping coefficient - '···fl.11 '! ,~ iii

Ji H-+H-t-flll+'-1++lW-+-H-+-H+H-l--+-+
1At+Hf-
.. h --H-lHI
and the natural frequency of the cylinder were .. ,.
determined in air and water by plucking excitation.
FORCE MEASUREMENTS 0.10 sec. == . '~
T;-C, F
Two cantilever-beam force transducers measured I C -
the instantaneous in-line (drag) and transverse I
(lift) forces. The strain gauges attached to the
beam emitted electric signals. Calibration of the
transducers was accomplished by hanging a load ~ FO,RCE .,.j',==t!l\ll,:t~':-~1=t1jt:tj=:tj=tj=tj=t1=t1=t:tjt:tj=
Ii liAJI1
from the center of the cylinder. This not only .. II
',I·'·· .!!:: '
established the level of the electrical output for a 0.10 sec. ,f-: +-H-++-
I ':;i::
known load, but also reaffirmed that the transducers Ii·: j II II
were positioned correctly.
FIG. 8 - SAMPLE DATA TRACES SHOWING THE
As with other elements of the experimental EVOLUTION OF THE IMPACT FORCE FOR
apparatus, these transducers had been used here VARIOUS RISE TIMES.
for more than 4 years. The details of their
construction may be found in Ref. 9. No changes
'.0
have been noted in the calibration of these
transducers since their installation.
Initial experimental efforts established that the
3.0 THEORET! CAL EXPERIMENTAL
transverse forces were small compared with the
in-line forces, and consequently measurement of
these forces was discontinued early in the study. /-{
The data were recorded with a high-speed tape , ,-,'-" \
\
,
and a strip-chart recorder. The strip-chart recorder \
operated at a speed of 200 mm/s, each division
representing O.OOS s. Fig. 8 is a typical example of
the data recorded. BUOYANCY

REDUCTION OF DATA z/o

Three force coefficients were defined In this 1.0 2.0 3.0

investigation. The first two are given by FIG. 9 - EVOLUTION OF THE BUOY ANT FORCE
AND THE THEORETICAL AND EXPERIMENTAL
C = 2F / (pDLU 2 ) . . . . . . . . . (12) VALUES OF THE IMPACT FORCE AS A FUNCTION
s m OF RELATIVE SUBMERGENCE.

34 SOCIETY OF PETROLEUM ENGINEERS JOURNAL

normalized impact force acting on the cylinder at of the system for a value of I ntT smaller than 0.9
the initial instants of the impact. This coefficient and attenuated for In t, values larger than 0.9. The
does not include the contribution of the dynamic second peak occurs only for IntT values larger than
response of the cylinder. 1.09. The amplification factor for this peak
The fluid force acting on the cylinder is increases from 0.B4 to about unity as Int, increases
amplified or attenuated by the dynamic response of from 1.09 to 1. 5.
the cylinder-transducer assembly. Thus, it was A comparison of Figs. 4 and Ba, 6 and Bb, and 7
neces~ary to subtract the inertial force acting on and Bc shows that the analysis can predict
the cylinder assembly from the recorded reaction accurately the evolution of the impact force and
force. First, the correct relationship between the help explain the differences in the initial value of
acceleration of the cylinder and the inertial force the impact force, if the rise time or Int, is known.
acting on the force transducer ~as established by Following the initial impact, the net force acting
plucking the cylinder and allowing it to freely on the cylinder begins to decrease, since the
vibrate in air. Then, the inertial force corresponding contribution of the added mass decreases. During
to the instantaneous acceleration was subtracted this stage of the flow, the cylinder undergoes
from the total force sensed by the transducer after damped oscillations at its natural frequency: As
the impact of the fluid surface on the cylinder. The the free surface rises further, the buoyant force
net fluid fNce was normalized as in Eq. 12 to increases and the separation effects begin to
yield the initial value of the slamming coefficient, create larger drag forces. The buoyancy-subtracted
C~. fluid force reaches its maximum at z/D values from
about 1.5 to 2. Even though the present flow
DISCUSSION OF THE RESUL TS situation, in which there is a free surface, cannot
be· compared directly with an impulsive flow about
Experiments with all the cylinders tested for all
a circular cylinder, the rise of the drag coefficient
Reynolds and Froude numbers yielded
to a maximum at z/D = 1.75 is much like the rise
CO ~3.17 ± 0.05· of the drag coefficient to a maximum at a relative
s fluid displacement of about 2 in impulsively started
Thus, the force transfer coefficient for slamming flow about a circular cy linder. 10 This rise in the
at the initial instants of impact is very close to drag coefficient results from the formation of two
its theoretical value of 1T. As shown previously, symmetrical vortexes behind the cylinder. As the
the maximum slamming force may be amplified by a motion continues, the vortexes become asymmetrical
factor of as much as 1.7 by the dynamic response and shed alternately.
of the structure. The maximum drag force was corrected for
Fig. 10 plots kx(t)/C~ as a function of Int, for buoyancy, normalized by 0.5pDLU~, and plotted
both the first and the second peak (see Figs. 8b as a function of the Froude number in Fig. 11.
and Bc), using wn values for both a 6-in.(l53-mm) Apparently, C h (see Eq. 13) decreases with in-
and a 3-in. (76-mm) cylinder. The drop in the creasing Froude number to a value of about unity
magnification factor for the first peak with and remains nearly constant for all Froude numbers
increasing rise in time is apparent. Also, the from 0.5 to about 2.
Experiments with cylinders having relative
effect of wn is confined to a narrow range of Int,
roughnesses of 1/100 and 1/50 have shown that
values smaller than about 0.5. The force acting on
the slamming coefficient, C~, remains unchanged.
the cylinder is amplified by the dynamic response
Roughness tended to increase the rise time and
decreased the force-amplification factor. Calcula-
kx(t)
C; tions for these coefficients were based on the
apparent diameters of the cylinders.

15 CONCLUSIONS
Theoretical and experimental investigation of

FOR SECONO PEAK ONLY zo

,
'A
. 4 in. (102 mm)
5 in. (127 mm)
J---
A

1.0
.. { A .. 6 in. (152 mm)
'0 • 8 in. (203 mm)
.- .-" ~.. 6-

" i.:.

" ~+:. ,.. •

...It: .•.•..o;...
• 6 eo....
·····a··!···················D
............ A •. !'•.... .......••...
AMI'UFICATION

05 Lo- - - - - - - - , 0 " " . 5 : - - - - - - -.....

FIG. 10 - AMPLIFICATION FACTOR AS A FUNCTION

1.'-0- - - - - - 1 . L .- - -
5
.. 10 1\ zo

OF Int, FOR TWO REPRESENTATIVE NATURAL FIG. 11 - DRAG-DOMINATED FORCE COEFFICIENT,

CIRCULAR FREQUENCIES. C h , AS A FUNCTION OF THE FROUDE NUMBER.

FEBRUARY 1979 3S
flow impact on circular cylinders warranted the 2. Kaplan, P. and Silbert, M. K.: "Impact Forces on
following conclusions. Platform Horizontal Members in the Splash Zone,"
paper OTC2498 presented at the Eighth Annual
1. . The dynamic response of the system is as Offshore Technology Conference, Houston, May
important as the impact force; one cannot be 3-6, 1976.
determined without accounting for the other. 3. Dalton, C. and Nash, J. J.: "Wave Slam on Horizontal
2. The initial value of the slamming coefficient Members of an Offshore Platform," paper OTC 2500
is essentially equal to its theoretical value of IT. presented at the Eighth Annual Offshore Technology
The system response may be amplified or attenuated, Conference, Houston, May 3-6, 1976 . .
depending on its dynamic characteristics. 4. Miller, B. L.: "Wave Slamming Loads on Horizontal
Circular Elements of Offshore Structures," paper
3. After impact, the cylinder undergoes damped
No. 5 presented at the Spring Meeting of the Royal
oscillations at its natural frequency. Institution of Naval Architects, London, April 5-7,
4. The buoyancy-corrected normalized force in 1977.
the drag-dominated region reaches a maximum at a 5. Faltinsen, 0., Kjaerland, 0., Nottveit, A;, and
relative fluid displacement of about 1. 75. Vinje, T.: "Water Impact Loads and· Dynamic
Subsequently, the shedding of vortexes and the Response of Horizontal Circular Cylinders in
Offshore Structures," paper OTC 2741 presented at
deceleration of flow reduces the maximum drag
the Ninth Annual Offshore Technology Conference,
coefficient to less than unity. Houston, May 2-5, 1977.
5. Since the rise time is not deterministic, the
6. Sarpkaya, T. and Isaacson, M. de St. Q.: Wave
sla~ing-force amplification or the dynamic Forces on Offshore Structures Theory and
response of the system should be analyzed using Application, in preparation.
the theoretical value of the slamming coefficient 7. Taylor, J. L.: "Some Hydrodynamical Inertia
and a rise-time distribution function. Coefficients," Philosophical Magazine 1930 Series
6. Roughness increases the rise time and tends 7, Vol. 9, 161-183.
to decrease the amplification factor. S. Thomson, W. T.: Theory of Vibration with AppZic{1.
tions, Prentice Hall, Inc., Englewood Cliffs, NJ
(1972).
ACKNOWLEDGMENTS
9. Sarpkaya, T:: "Vortex Shedding and Resistance in
The author thanks Richard A Post, Neil J. Harmonic Flow about Smooth and Rough Circular
Collins, and Jack McKay for their help and Cylinders. at High Reynolds Numbers," Tech. Report
No. NPS-59SL76021, Naval Postgraduate School,
assistance with the experiments. Monterey, CA (Feb. 1976).
10. Sarpkaya, T.: "Impulsive Flow About a Circular
REFERENCE Cylinder," Tech. Report No. NPS-69SL-7S-00S,
1. Szebehely, V. G.: "Hydrodynamic Impact," Applied Naval Postgraduate School, Monterey, CA (March
Mechanics Reviews (1959) Vol. 12, No.5, 297-300. 1975).
***

36 SOCIETY OF PETROLEUM ENGINEERS JOURNAL