You are on page 1of 8

Proceedzngs of the Fourth (1994) InternatIOnal Offshore and Polar Engzneenng Conference

Osaka, Japan, Aprill0-lS, 1994


Copynght © 1994 by The International Society of Offshore and Polar Engineers
ISBN J-880653-10-9 (Set); ISBN J-880653-]3-3 (Vol.llI)

Added Mass and Damping on an Oscillating Surface-Piercing Circular


Column with a Circular Footing
Jin S. Chung
Colorado School of Mines
Golden, Colorado, USA

ABSTRACT predictions of the added mass by potential theory are 10-20% lower than
the experimental values at the model submergences tested. Kudo and
Added mass and damping on a vertical, surface-piercing, circular Kinoshita (1981) conducted an experiment for the vertical oscillation of
cylinder or column with a submerged sharp-cornered circular similar models of smaller size in infinite depth and also reported the
footing, oscillating in water of finite and infinite depths are measured underestimates of the coefficients by their source distribution method.
using a planar motion mechanism. The model is made of aluminum,
and the outer diameters of the column and footing are 26.67 em and Most floating semisubmersible structures for offshore petroleum
60.96 cm, and their lengths are 90.17 em and 22.86 cm, respectively. drilling and production operations consist of members of circular,
The bodies were forced to oscillate sinusoidally with small rectangular and other cross sections. Their submerged members oscillate
amplitudes, for several submergences below a free surface. The near the free surface. For submerged bodies oscillating in an infinite
added-mass and wave-damping coefficients are shown to be fluid, the added-mass coefficients are constants, and the wave damping
influenced strongly by the free-surface effect and are presented as a is zero. But as the submerged body oscillates close to the free surface,
function of water depth, frequency and direction of oscillation and the free-surface effect can influence the added-mass and wave-damping
of depth of submergence from the free surface. For the vertical coefficient values greatly, as a function of the body geometry, frequency
oscillation close to the free surface, negative added mass values are and direction of oscillations, and water depth. These frequency-
measured, and the predictions of the added mass by a 3-D diffraction dependent coefficient values provide better accuracy for computing the
theory are 10-20% lower than the experimental values at the model motion of a floating structure and forces on its submerged structural
submergences tested. The experimental added mass coefficient values members.
for the horizontal oscillation and the wave damping coefficients for
the vertical oscillation in finite depth differ more than 100% from the As a contribution to this matter, with emphasis on practical
corresponding 3-D theory prediction at a certain at-sea operational applications, a series of experiments were carried out with the harmonic
frequency range. This set of data provides further experimental oscillations of a vertical surface-piercing circular cylInder with a sharp-
information for the improvement of theoretical predictions. ' cornered circular footing (e.g. similar to a Sedco-135 series, 3-leg
semisubmersible structure) in still water having a free surface. The
KEY WORDS: Vertical circular cyl inder, circular footing, horizontal and experiments are conducted at several submerged distances over the
vertical oscillations, finite depth, added mass, damping, negative added frequency range from the free surface for both finite and infinite depths.
mass The model geometry, submergence and frequency range represent closely
the actual ocean operation of the 3-leg Sedco-135 series semisubmersible
Unit conversion: 1 in = 2.54 cm. structures. The corresponding added-mass and wave-damping coefficients
for both vertical and horizontal oscillations were determined, including
INTRODUCTION the effects of the free surface and finite depth. In-phase and out-<lf-phase
forces, which are linear with the oscillation amplitude, were measured
Recently, 3-D (three-dimensional) potential theories and computational by a planar-motion mechanism. The motion amplitudes are kept small
methods have made much progress in contributing to the theoretical with respect to the dimensions of the cross sections, and the generated
prediction of hydrodynamic forces and motions of floating structures. waves have amplitudes that are small with respect to the wavelength.
However, the accuracy of theoretical predictions needs to be confirmed The measured added-mass coefficients are presented in a form suitable
by experimental results. Previously, Miao, Liu and Chung (1985) with the hydrodynamic force equation (Chung, 1976) within the range of
showed for a vertical circular cylinder or column with a circular footing the experimental variables. As the present experiment was conducted for
the purpose of specific engineering applications, data points at a few
that the prediction by a 3-D diffraction theory was more than 100%
theoretically interesting submergences and frequencies were not taken.
different at a certain frequency range of horizontal oscillation from the
corresponding experimental values. Also for the vertical oscillation, the

182
MODEL AND EXPERIMENTS horizontal and vertical directions at the beginning of testing over the
range of amplitudes and frequency at the shallowest submergence tested
General Setup (s' = 1.25) (Fig. 2 and Table 1). The s' is defined as s' = s/a, where
s is the distance to the half height of the footing from the still-water
Experiments were conducted in a towing tank, whose dimensions are surface and a is the half-height of the footing. Linear amplitudes in each
122 m in length and 7.62 m in width. Water depth was h =4.57 m. mode of oscillation of each model were chosen over the frequency range,
About halfway along its length was a planar motion mechanism (PMM) using the in-phase force curves as a functIOn of oscillation amplitudes.
(Goodman, 1968) installed on a stiff towing carriage. Braces were A typical in-phase force-to-amplitude relationship for the circular cross
section is established, and the measured in-phase forces are presented in
installed between the PMM assembly and the towing carriage for
Fig. 2a as a function of horizontal oscillation amplitudes, 1J, and in Fig.
stiffness, and the entire carriage was lifted off its wheels and supported
2b as a function of vertical oscillation amplItudes, .i- At w' = 0.3 and
on blocks to provide a rigid base during testing. A sonic probe measured
0.4 where w' = w2a/g, the results of the vertical in-phase force,~, show
the outgoing waves at a distance of about 2 m from the axis of the
model.
some degree of nonlinearity at r = 1.016 cm. However, the largest
values among the oscillation amplitudes that assure the linearity over the
frequency range (except these 2 frequencies) were determined as r =
The forced harmonic oscillation experiments were conducted for a
1:016 cm (0.4 in) for vertical oscillations, and 1J = 2.032 cm (0.8 in) for
vertical surface-piercing circular cylinder with a sharp-cornered circular
the horizontal oscillations (Table 1). These amplitudes, taken at s' =
footing (e.g. similar to the Sedco-135 series semisubmersibles). The
1.25, were used to measure and analyze the forces at all model
joint between the cylinder and the footing and the footing itself are
submergences s' ~ 1.25.
sharp-cornered (Figs. la and Ib). This model is commonly used for
members of semisubmersible (floating) structures. To minimize possible
Simulation of Infinite Water and Finite Depth
scale effects such as those associated with cross-flow drag at low
Reynolds numbers, as well as measurement errors, the model was made
The damping force components, to produce no wave damping, were
of alummum, as large as possible, yet compatible with the tank
measured by oscillating the models in the horizontal and vertical
dimensions: the outer diameters of the column and footing are 26.67 cm
(10.5 in) and 60.96 cm (24 in), respectively, and their corresponding directions, respectively, over the frequency range in the simulated
lengths are 90.17 cm (35.5 in) aiJd 22.86 cm (9 in) (Fig. Ib and Table infinite water. The infinite water was simulated at the submergence of
1). s' = 8 for the vertical oscillation, for which negligible outgoing waves
were measured. The s' is defined as s' = s/a, where S IS the distance
Force measurements were made by means of standard 1O.16-cm to the half height of the footmg from the still-water surface and a is the
modular force gages, which are described in Goodman (1968). These half height of the footing. The tank water depth was h' = 33.8 The h'
gages are essentially flexure boxes machined out of a solid block of is defined as h' = h/a, where h is the water depth. The simulation at s'
ARMCO 17-4 PH stainless steel. The deflection of the flexures is a > 8 was difficult, because the supports joining the top of the vertical
linear function of applied force and IS sensed by a variable reluctance cylinder of the model to the PMM would not be absolutely rIgid when
gage. The force was shared by a gage on the top center of the vertical subjected to large forces produced by the model durmg the horizontal
surface-piercing cylinder. To provide accurate and contmuouf, oscillatIOns at s' > 8. The s' = 8 submergence came close to the
measurements of the oscillation amplitudes of the model, a cable-pulley-, simulation of infinite water, but a small free-surface effect still could be
potentiometer arrangement was used. The electrical signal, which was detected at the lower end of the frequency range in both the added-mass
calibrated as a function of amplitude, was recorded on an X-Yplotter. and damping coefficients (Figs. 3-4). However, the dampmg at s' =
8.leveled off for infinite depth and slightly increased for finite depth, at
Measurements higher frequencies, at which negligible outgoing waves were measured
for the vertical oscillation.
The horizontal and vertical oscillations were produced directly by the
planar motion mechanism (PMM) with an adapter, which allowed either Finite depth was simulated with a false bottom made of marine
an horizontal or vertical mode. For the vertical oscillation, the plywood, located at h' = 8.0 where h was 36 in (91.44 cm).
amplitudes of the PMM oscillator and of the models were identical.
However, very small bending of the model was recorded during the Frequency and Submergence
horizontal oscillations, which increased slightly the model amplitudes as
The nondimensional oscillation frequency range selected for the
compared to the oscillator amplitudes. The model amplitudes were used
present experiment series represents closely that of the Sedco-135 series,
in reducing the force data to the coefficients. The waves reflected by the
3-leg semisubmersible structures for their at-sea operations: w' =
tank wall were carefully monitored to avoid their effects on the measured
0.05-2.0 where w' = w2a/g. The wave period range of interest for
force data. When the model was oscillated in air over the frequency
ocean operations can be 5 ~ T ~ 25 s.
range to obtain tare force measurements, the amplitudes of the oscillator
and models were identical for both horizontal and vertical modes.
Submergences of the half height of the footing below the still-water
surface were selected as s' = 1.25, 2.0, and 8.0 for infinite depth and
. Measured signals coming from the force gages were processed
s' = 4.0 and 6.0 for the finite depth of h' = 8. These sub mergences are
automatically onboard the carriage by the signal processor (Goodman,
suitable for practical applications to the at-sea operations of the Sedco-
1968) at the end of each test to get in-phase and out-of-phase force
135 series structures. The test series were programmed to emphasize the
compo~ents that correspond to the added mass and damping,
effects of the free surface and water depth on the coefficients as a
respectl~ely. Wave-damping force components were obtained by
function of the model submergence over practical ranges of frequency
subtractmg the out-of-phase forces measured ill simulated infinite water
and submergence. Coefficients for s' = 8 were affected slightly at the
from the out-of-phase forces measured at each submergence. lower end of the frequency range by the free surface, according to the
data (Figs. 3 -4).
Linearity of Forces with Oscillation Amplitudes
Added Mass and Damping Coefficients
In order to establish the linearity of the measured forces with the
oscillation amplitudes, each model was oscillated separately in the The nondimensional, added-mass coefficients a.u and wave-damping

183
coefficients bJl are obtained after removing the in-air tares and the flow interactions at the column-footing joint.
damping caused by viscosity:
Submergence and Oscillation Frequency
For horizontal oscillation (j = 2):
For a fluid that is bounded by a free surface, the coefficients vary as
ayy = Y,111Ual 21J - m'lm) a function of oscillation frequency and submergence of the model For
byy = Y/11Ual1J - b"yylm) an infinite fluid, the added-mass coefficients are constants and are
independent of the oscillation frequencies for small amplitudes; the
for vertical oscillation (j = 3): corresponding wave-damping coefficients are zero.

azz = Z/ml.oit - m'lm) As the submergence of the model becomes shallower, the free-surface
b" = Z/mwt - b"dm) effect gets stronger and is more pronounced at the lower end of the
frequency range, at which the maximum and minimum values of added-
where Y" Z, = horizontal and vertical amplitudes, respectively, of the in- mass coefficients are found. This lower end of the frequency range is of
phase force, Yo, Zo = horizontal and vertical amplitudes, respectively, of interest for practical operations of the floating ocean structures. A
the out-of-phase force, m = displaced-water mass (m =7rpr(s-a) + decrease in the model submergence increases the maximum values of the
7rpR2(2a», m' = tare mass of the model obtained in air test, a = the added-mass and wave-damping coefficients and also reduces the
half height of circular footing, r = column radiUS, R = footing radius, minimum values of the added-mass coefficients.
b"yy = horizontal damping coefficients obtained at s' = 8 (b"yy = Y,JW1J),
b"zz = vertical damping coefficients obtained at s' = 8 (b"zz = Z,Jwf), 1J For the vertical oscillations very close to the free surface with infinite
= amplitude of horizontal oscillation, and t = amplitude of vertical depth (s' = 1.25), the experiments show that the minimum value of the
oscillation. added-mass coefficients becomes negative (Fig. 3), to which much
attention was paid dunng the testing to the free-surface effect and the
RESULTS AND DISCUSSIONS negative added mass in particular. The negative experimental pomt was
measured for s' = 1.25, and the curve trend shows that it occurs at w'
Nondimensional added-mass and wave-damping coefficients are = 0.6-1.3. The experimental values of the negative added mass also
presented m Figs. 3-4 for infinite depth and m Figs. 5-6 for fimte for the vertical oscillation of the horizontal cylinder of square cross
depth (h '= 8) The consistency of the experiments is satisfactory. section were previously reported by Chung (1977). It is speculated that
the dip or negative value at w' = 1.0 in the added mass curve may be
Direction of Oscillations caused by the interactions among the waves generated successively by the
shallow-submerged footing during the vertical oSCillatIOn. In order to
The measurements show that peak coefficient values are much larger observe the surface wave interactions, the model was vertically oscillated
for the vertical than for the horizontal oscillations. As expected from the at w' = 0.565 and s' = 1.25 with t = 1.016 cm (0.4 in) where the
shape of the model, added mass coeffiCients for the vertical oscillation added mass is negative and linearity was established (Fig. 8). Splashing
(Fig. 3) are much larger over the frequency range than those for the of surface waves started on the surface-piercing column near the column-
honzontal oscillation (Fig. 4). The free-surface effect is more footing joint. Progressively, ring waves were formed on the top surface
pronounced for the vertical oscillation. of the footing, starting from the column, and wave crest rings were
visible on top of the footing surface and around the footing, and stayed
The measurements at s' = 1.25 show that the added mass coefficients as if standing waves. The photo shows clearly the flow separation from
for the vertical oscillation become negative at the frequency range of w ' the sharp corner of the footing top.
= 0.6-1.3. The measurements at both s· = 1.25 and 2.0 show a
strong free-surface effect. At the deep submergence (s' = 8) the added At s' = 8.0, the added-mass coefficients approach their respective
mass coefficients approach nearly a constant of a" = 0.9 at higher constant values at higher frequencies, and the corresponding wave
frequencies. damping is small. Although the s' = 8.0 submergence was used to
simulate mfinite depth, the added mass coefficients measured show a
For the horizontal oscillation (Fig. 4), the free-surface effect is strong small presence of the free-surface effect at the lower frequency range,
at s' = 1.25 and 2.0 and negligibly small at s' = 8.0, where the added and the wave damping appears to become very small beyond w' = 2.0.
mass coefficient stays nearly constant with ayy = 0.6 at higher At higher frequencies near w' = 2.0, the added-mass coefficients of both
frequencies. For the horizontal oscillations, vortiCes were observed models approach constants for the respective submergences and both
around the column and the footing. Possible free-surface effect on the directions of oscillation, while the wave-damping coefficients seem to
added mass and damping due to the surface-piercing column were not disappear well beyond w' = 2.0.
established because of lack of the experimental data points at the lower
end of the frequency. The column effects are further investigated by Finite-Depth Effect
Chung (1994).
Both added-mass and wave-damping coefficients for the finite depth
As compared to a theoretical prediction (Miao, Liu and Chung, 1985), increase (Fig. 5 - 6) as the model oscillates closer to the bottom for both
the experimental results of added mass for the vertical oscillation are 10- horizontal and vertical oscillations. For both directions of oscillations,
20% larger over the frequency range than the theoretical prediction (Fig. added mass coefficients with sharp dip points were measured near w' =
7). The plotting error in the experimental damping was noticed in Miao 0.5, and more distinct for the horizontal oscillation. For the
et al. (1985) and the corrected values are replotted in Fig. 7. The nondimensional frequency, the definition of w' = w2a/g is used m Figs.
theoretical prediction for the horizontal oscilJation shows a. large 5-6.
discrepancy from the experiment results over the frequency range. Near
w' = 0.5 the discrepancy is very large. As to the wave damping, the The 3-D theory (Miao, Liu and Chung, 1985), which was replotted
comparison shows that the discrepancy is large for the vertical oscillation more clearly by Chung (1994a), tends to overestimate the added mass
over the frequency range for the experimental condition of h' = 8.0 and greatly over w' = 0.3-0.8 and w' > 1.3 for the finite depth of h' =
s' = 6.0. This discrepancy may be caused by the flow separation and 3- 6.0 for the horizontal oscillations This discrepancy is more pronounced
D ring vortices from the sharp corners of the footing surface and by the for the horizontal OSCillation at s' = 4.0 and 6.0 of finite depth where

184
large dip points are measured at w' = 0.5 at both s' = 4.0 and 6.0. For Miao, GP, Liu, YZ, and Chung, JS (1985). "A Note on the Added Mass
the vertical oscillations the theory underestimates the added mass by 10- and Wave Damping of a Column with Circular Footing," J Energy
20% over the frequency range for both finite and infinite water depths. Resources Tech, ASME, Vol 107, pp 235-237.
As speculated previously, this discrepancy may be caused by the flow
separation from the sharp corners of the footing surface and the wave
interactions at and 3-D vortices from the sharp-cornered footing surface. Column length, L (cm) 90.17
Measured wave damping is small, but is much larger for the vertical Column diameter, 2r (cm) 26.67
oscillation than the theoretical prediction (Fig. 7). The variation of the Footing height, 2a (cm) 22.86
coefficients at the low frequencies is probably influenced by the free Footing diameter, 2R (cm) 60.96
surface. Linear oscillation amplitudes
used for the test series:
CONCLUSIONS AND RECOMMENDATIONS Horizontal amplitude, '1/ (cm) 2.032
Vertical amplitude, ~ (cm) 1.016
For the vertical oscillation close to the free surface, negative added Finite depth, h (cm) 91.44
mass values are measured, and the predictions of the added mass by
potential theory are 10-20% lower at the model submergences tested than Table 1 Principal characteristics of the model and osciilation amplitudes
the experimental values. The experimental added-mass coefficient values
for horizontal oscillation in finite depth differ more than 100% from the
corresponding 3-D potential theory prediction at a certain at-sea
operational frequency range. Measured wave damping is small, but much
larger for the vertical oscillation than the theoretical prediction (FIg. 7).
The model is made large to minimize potential errors from the
installation and measurements. This set of data provides further
experimental information for the use of correct coefficient values and the
improvement of theoretical predictions.

Despite recent progress in 3-D potential theories and computational


methods for the theoretical prediction of hydrodynamic forces and
motions of floating structures, the present experimental results suggest
that the accuracy of theoretical predictions needs to be confirmed by
experimental results. The discrepancy between potential theories and the
experiment may be caused partly by the flow separation from the sharp
corners of the footing surface and the wave interactions at and 3-D
vortices from the sharp-cornered footing surface for the vertical
oscillations. Also for the horizontal oscillations, vortices were observed
around the column and the footing. Potential theory does not account for
the flow separation, and its predictions may not be very accurate for the
sharp-cornered model.

The free-surface effect of the column could not be discussed, due to


lack of data points taken at the lower frequencies. The free-surface effect Fig. la Model of vertical circular cylinder with circular footing rigged
is further discussed by Chung (1994b) with another experiment of a to PMM system over false bottom
square cross-sectional column.

REFERENCES

Chung, IS (1976). "Motion of a Floating Structure in Water of Uniform


r-- 2r = 26 67 em

Depth," J Hydronautics, AIAA, Vol 10, pp 65-73.

Chung, JS (1977). "Forces on Submerged Cylinders Oscillating near a


Free Surface," J Hydronautics, AIAA, pp 100-106.
L
T
= 9017 em

Chung, JS (1994a). "Added Mass and Damping on an Oscillating t


Surface-Piercing Circular Column with a Circular Footing," Int J s
• Offshore and Polar Eng, ISOPE, Vol 4, No 1.
Chung, JS (1994b). "Added Mass and Damping on an Oscillating
Surface-Piercing Square Column with a Horizontal Square Cylinder," to a = 11 43 em
be published.

h 2R = 6096 em
Goodman, A (1968). "Description and Operation of Planar Motion
Mechanism System," Hydronautics Inc, Tech Manual 754-1.
I
!
Kudo, K, and Kinoshita, K (1981). "Fundamental Study of Vortex-
Making Damping Forces and the Motions of a Column with Footing" (in
Japanese), J Soc Naval Arch, SNAJ, Japan, Vol 150, pp 285-293. Fig. Ib Schematic of model of vertical circular cylinder with circular
footing

185
100

(b)
(a) 30

()
"v.
II

:c 20
I\j-

ai
g 0
u
>-- u..
CD
ai (/)
U CO
0 .r::.
u.. c..
CD
(/)
E
CO
50 "iii
.r::. u
c.. ;:;
C Q;
> 10

c ,,0
0 ,~
~
ul
0
:r:
w' = 0 51. w' =0 3
- - - --- -4;Jr
-;- - ==-.= - - -1::. ... w' = 04
- - - - -0,,- w' = 0.2
w' = 01
o
o 0.2 04 06 08 10

Vertical OSCillation Amplitude, (' (m)

Fig. 2b Linearity of the measured in-phase force as a function of


vertical oscillation amplitudes of the model at s' = 1.25
o 0.2 0.4 06 08 1.0

HOrizontal Oscillation Amplitude, f] (In)

Fig. 2a Linearity of measured in-phase force as a function of horizontal


oscillation amplitudes of model at s' =
1.25

186
3-r------------------------------~ 3-r--~r_------------------------~

ro~ oClll

~
.,ui'" is.
::E
"0
.,
E
Q)
0
"0
"0
<I:
.,~
~

o 2 o 2
w' = w2alg w' = w2alg

Fig. 3 Added-mass and wave-damping coefficients for vertical oscillation in infinite depth

1.0-r----------------------------~
s' = sla
h' = hla

h' = 00

s'= 1.25

s' = 8.0

$'= 1.25

o 2 o 2
w' = w2alg w' = w2alg

Fig. 4 Added-mass and wave-damping coefficients for horizontal oscillation in infinite depth

187
2~-----------------------------.
h' = 8.0

s' = sla

s' = 6.0

s' = 4.0

w' = w2a/g

0.5
:..
h' = 80
:..
~

ci s' = sla
c:
0 '0.
0 2 E
to
0
CD
s' = 4.0
>
to
~
0.5 s' = 6.0
0.0
~
~ 0 2
ci
c:
0.
E
to
Cl Fig. 6 Added-mass and wave-damping coefficients for horizontal
CD
> s' = 40 oscillation in finite depth (h' = 8)
'"
~
s' = 6.0
0.0
0 2

~
Fig. 5 Added-mass and wave-damping coefficients for vertical
oscillation in finite depth (h' = 8)

s' = 6, h' = 8 s' = 8, h' = =

Experiment Experiment
Theory Theory

0.5

1.0

o 0.5 1.0 1 5
05

(b) Wave Damping

ria = 1.167
R/r = 2.286
(a) Added Mass

Fig. 7 Comparison of itheoretical predictions with experimental results (Miao, Liu and Chung, 1985)

188
Fig. 8 Flow separation from sharp corner of footing top and interactions
= =
of surface ring waves on and around footing: w' 0.565 and s' 1.25
r
and = 1.016 cm

189

You might also like