M. B.

Bryndum
Hydrodynamic Forces on
V. Jacobsen
Pipelines: Model Tests
An extensive model test program on the hydrodynamicforces on a submarine pipeline
Danish Hydraulic Institute, resting on the sea bottom and exposed to steady current, regular waves, combined
Horsholm, Denmark steady current and regular waves, irregular waves, and combined steady current and
irregular waves has been performed. The hydrodynamic forces in both the in-line
and the cross-flow directions have been analyzed using three different methods, i.e.,
D. T. Tsahalis least-squares-fit analysis based on Morison-type equations, Fourier analysis, and
maximum force analysis. The force coefficients associated with each method have
Department of Chemical Engineering, been determined for a wide range of environmental conditions. The results of the
University of Patras, tests are presented in terms of the calculated force coefficients and their dependence
Patra, Greece
on various nondimensional parameters is discussed. Furthermore, comparisons with
other test data are presented.

Introduction
The aim of on-bottom stability design of submarine pipelines called carriage technique, which combines the advantages and
is to ensure their structural integrity when exposed to envi- flexibility of a medium size facility with the capability of car-
ronmental forces. Traditionally this has been achieved by re- rying out tests at full scale or near full scale. The model as
quiring that the pipeline should remain stable, i.e., no such is composed of the model pipe itself and a part of the
movement when exposed to the environmental forces under seabed, i.e., a flat plate sufficiently long to allow a correct
design conditions. However, various pipeline design codes al- oscillatory boundary layer to develop. The model is suspended
low certain limited movements of the pipeline that do not lead vertically from a carriage which runs on rails mounted on top
to interference with adjacent objects or overstressing of the of the flume walls. The carriage can be driven with any pre-
pipe. scribed oscillatory motion and can thus move the model relative
Both design philosophies require accurate and reliable pre- to the water, reproducing the near seabed wave-induced hor-
dictions of the forces involved. However, a design based on izontal water particle motions. A return flow in the flume
the no-movement criterion requires only knowledge of the reproduces a steady current. The model pipe and seabed are
worst combination of forces, whereas a design in which lateral fixed to each other and the gap between the pipe and the seabed
movements are allowed requires information on the time var- is sealed. The model is shown in principle in Fig. 1.
iation of the forces over the entire range of relevant environ-
mental conditions. Model Setup. The flume used for the model tests has a
The hydrodynamic model tests described are part of the test section 35 m long and 3 m wide with a maximum water
American Gas Association's long-term research effort to im- depth of 0.8 m, see Fig. 2. The maximum rate of flow is 1.5
prove existing and develop new design methods for the on- m3/s, which for a 0.8-m water depth and a width of 2.7 m
bottom stability of submarine pipelines. The test program and gives a cross-sectional average velocity of 0.69 m/s.
the data analyses were planned and executed in such a way Pipe and seabed were mounted in vertical position close to
that the resulting data could be equally applicable to either one of the flume walls. To prevent deflections of the 8-m long
design philosophy. model seabed, a guiding rail system was mounted at the bottom
This paper presents the model tests and the results in terms of the flume. Furthermore, a fixed intermediate wall was in-
of the force coefficients. In a separate paper, reference [10], stalled just behind the moving seabed extending through the
the hydrodynamic forces due to irregular wave and current full length of the flume. This wall served a double purpose;
action are discussed, and a new analytical model providing it eliminated secondary flows behind the model seabed, and
accurate force predictions is presented.
PLAN OP MODEL S E T - U P

Model Tests
FIXED BOUNDARIES
Testing Technique. The testing technique applied is the so- FLUME WALLS
MODEL PIPE
Contributed by the OMAE Division and presented at the 7th International MODEL 5EABEO
Symposium and Exhibit on Offshore Mechanics and Arctic Engineering, Hous- r~\ SEALING
ton, Texas, February 7-12, 1988, of THE AMERICAN SOCIETY OF MECHANICAL W////////////////W///////^^^
ENGINEERS. Manuscript received by the OMAE Division, March 20,1990; revised
manuscript received June 11, 1991. Associate Technical Editor:
D. Myrhaug. Fig. 1 Model setup in principle

Journal of Offshore Mechanics and Arctic Engineering NOVEMBER 1992, Vol. 114 / 231

Copyright 1992 by ASME

Downloaded 15 Jan 2013 to 223.27.129.246. Redistribution subject to ASME license or copyright; see http://www.asme.org/terms/Terms_Use.cfm
 CARRIAGE QN RAILS (WEia.

mill IYTTP I I I I I ITI I I I I I I rtn

ELEVATION GLASS SECTON

3 5 m TEST SECTKJN

SECTION B - B

Fig. 2 Current flume and oscillating carriage

helped generate the desired steady current profile when pro-

vided with roughness panels. The surface wave disturbances
created in the flume when moving the model were eliminated
by means of a large horizontal aluminum plate 2.5 m wide by Fig. 3 Cross section in flume with pipe suspended from carriage
7.5 m long. A cross section of the model setup is shown in
Fig. 3.
ducer. The near seabed undisturbed wave-induced particle ve-
Model Pipes. The model pipes were composed of three locities and accelerations, which were produced by the carriage
parts. The central measuring section was suspended in the force motions, were measured directly by means of a tachometer
transducer, whereas both ends were fixed directly to the seabed and an accelerometer mounted on the carriage.
and acted as "dummies" eliminating end and other model In addition to these basic measurements, the steady current
effects on the forces measured by the live section (see Fig. 4). in the flume was measured using a reference current meter
Two pipe models were used, having external diameters of placed upstream of the pipe in the centerline of the flume, and
200 mm and 400 mm. Each pipe was 750 mm long, including a flow meter mounted in the flume recirculation system. For
the gaps between different sections. The seabed and both model future analyses and improved understanding of the relation
pipes were provided with surface roughness which was varied between the flow around the pipe and the induced forces, an
during the course of the test program. Three distinctly different array of bi-directional current sensors were placed in the vi-
surface roughnesses were applied: cinity of the pipe. The position of the current sensors are shown
in Fig. 5.
Fine k/D=\0~3
Medium k/D =10 - 2 Model Test Program
Rough k/D = 5-\Q-
The test program was divided into two parts:
Instrumentation. The hydrodynamic forces on the instru-
mented pipe segment were measured directly in the in-line and 1 The Main Test Matrix comprised approximately 500 in-
cross-flow directions using a two-component shear force trans- dividual tests covering the following environmental conditions:

KC = U T/D regular
a -= acceleration waves t = time
ahb, --= Fourier coefficients KC = Um0 Tp/D irregular U = velocity
Ca == added mass coefficient waves Uc = current velocity
cD == drag coefficient k = hydraulic roughness of pipe Uc = current velocity averaged
CH == nondimensional horizontal on seabed over pipe diameter
force kb = hydraulic roughness of Um0 = spectrum derived significant
Q --= Fourier coefficient seabed bottom velocity
cL --= lift coefficient k* = pipe roughness ratio ( = k/ Uw = maximum wave velocity
CM -= inertia coefficient D) V= mean velocity outside
Cy --= nondimensional lift force kt = seabed roughness ratio boundary layer
D --= pipe diameter ( = kb/D) W = weight per unit length
F --= force MME = structural mass y = vertical distance from
FD == drag force per unit length N = total no. of time steps or seabed
F = = total horizontal force per Fourier coefficients a = current ratio
unit length Re = Reynolds no. a = Uc/Uw regular waves
F, ~-= inertia force per unit length Re = Uc D/v steady cur- a = Uc/Um0 irregular
FL '-= lift force per unit length rent waves
Fv == lift force per unit length Re = Uw D/v waves v = kinematic viscosity
-FMEAS = = measured force T - wave period of regular p = density of water
-FpRED = = predicted force waves <>
/ = phase
I == time step, counter Tp = peak period of bottom ve- co = cyclic frequency
KC == Keulegan-Carpenter no. locity spectrum co = 2ir/T

232 / Vol. 114, NOVEMBER 1992 Transactions of the ASME

Downloaded 15 Jan 2013 to 223.27.129.246. Redistribution subject to ASME license or copyright; see http://www.asme.org/terms/Terms_Use.cfm
 Table 1 Overview of parameter range for the main test matrix
Parameter Definition Test range
*ipe SECTION DLMMY PPE SECTIC
^ Keulegan-Carpenter
no.
B A S E PLATE
MODEL SEABED
u T 3-160
Regular waves D
ELASTIC 5EAUNO
ELASTIC 5EAL1NQ
U,o T p 1 10-70
FORCE TRANSDUCER FORCE TRANSDUCER
Irregular waves D
AOCELERQMETER ACCELEROMETER

m Current ratio

DUMMY PIPE SECTION

Regular waves 0-1.6
DUMMY PIPE SECTION

fiUlD6HAlL_
s\sSJSSS77? 7
Irregular waves 0-1.2
Fig. 4 Model pipes U,o

Reynolds no. 10=

STEADY CURRENT Uc D
Steady current Re- 0.3-2.4
* V

WAVE MOTION

FLOW VELOCITY Regular waves Re = uw D 0.5-3.6

MEASUREMENT MODEL PIPE S U S P E N V
POSITION DEO IN X- Y FORCE
TRANSDUCER U,o D
Irregular waves 0.7-2.5

0 25 D Pipe roughness ratio Ar*=- 10" 3 -5-10" 2

0.25 K D
D

Seabed roughness ratio kb 10~ 3 -5-10~ 2

~D
Fig. 5 Instrumentation in principle

9
Steady current Table 2 Test matrix for regular waves with and without cur-
9
Regular waves rent, D = 0.2 m, k/D = NT 3 , NT 2 , 5 10 2
9
Regular waves and steady current
Current ratio: a
* Irregular waves
9
Irregular waves and steady current KC- 0.0 0.16 0.32 0.48 0.64 0.80 0.96 1.20 1.60
The influence of the most significant nondimensional param- no.
eters, i.e., the KC-number, the current ratio, and the pipe 2.5
4.5
roughness ratio, on the hydrodynamic forces has been deter- 5.0
mined from these tests. 10 1 1 I 1
2 The Supplementary Test Program also comprised ap- 12
proximately 500 tests. The purpose of these tests was to in- 15 1 1 I 1
17
vestigate the influence of certain parameters which have 20 1 1 1
secondary influence on the hydrodynamic forces. These are 25
the seabed roughness ratio and the wave irregularity in natural 30 1 1 1
wave trains. In addition, the effect of working in reduced scale 40 1 1
50 1 1
has been evaluated. 55
60
An overview of the parameter ranges covered in the tests is 65
given in Table 1, and a more detailed test matrix for the main 70
75
test program is given in Table 2. 80
100
120
Methods of Analysis, Regular Waves/Steady Current 140
The objective of the different analyses applied was to es- 160
tablish nondimensional force coefficients which in combina- (ff
1" indicates that one or more tests have been performed
tion with analytical expressions or numerical methods would
make it possible to reproduce correctly the hydrodynamic forces
experienced by prototype pipelines. The methods applied dur- ficients for drag, inertia, and lift, commonly used for calcu-
ing the test program were: lating hydrodynamic forces on pipelines, i.e., the Morison type
of force formulation, have been determined using the following
(a) least-squares-fit (LSF) method based on a Morison type expressions describing the horizontal or in-line force and the
of force equation for both in-line and lift forces yielding tra- lift force:
ditional force coefficients for drag, inertia, and lift;
(b) decomposition of the measured forces into nine-compo- FH(t)=^pDCDU{t) 11/(0 I + J pD*CMa(t) (1)
nent Fourier series;
(c) extreme force analysis yielding maximum coefficients for
each half-wave. FL(t)=-PDCLU'(t) (2)

Least-Squares-Fit Analysis. The values of the force coef- where U(t) = Uw(t) + Uc is the total velocity and a(t) is

Journal of Offshore Mechanics and Arctic Engineering NOVEMBER 1992, Vol. 114 / 233

Downloaded 15 Jan 2013 to 223.27.129.246. Redistribution subject to ASME license or copyright; see http://www.asme.org/terms/Terms_Use.cfm
 the acceleration. CD, CMand CL are the traditional coefficients dominated by the inertia part. Therefore, the drag was used,
for drag, inertia, and lift, respectively. The force coefficients obtained by subtracting the inertia force from the measured
were found through a least-squares-fit between the measured horizontal force
force components and those predicted by Eqs. (1) and (2).
Due to the testing technique the in-line force formulation is FD(t)=FH(t)~^pD2Ca-a(t) (9)
slightly modified because the model and not the water is ac- 4
celerated. Equation (1) can be written in more detail as where Ca = 2.29, determined from potential flow theory, was
used in the Fourier decomposition.
FH(t) = l-PDCDU(t) I U(t) I + J PD
2
Caar{t) The force coefficients were determined using the following
expressions:
+ ?-PD2aM) (3) CH(t)=FD(t)/\pDU2w
4
9
where aw is the acceleration of the water, which in nature is
caused by the pressure gradient from the sloping surface ele- = CHO+YJ Cm-cos i(wt-cj>Hi) (10)
vation. ar is the relative acceleration between pipe and water, /=i

and ar equals aw for a fixed pipe. C = CM- 1 is the added

mass coefficient. In the present model tests, Eq. (3) becomes Cv(t)=FL(t)/^PDUl
9

FH(t) =~ PDCDU{t) I U(t) I +^pI?Caar(t) = CVo+^CVi-cos i(o)t-<pvi) (11)

;'=1
+ MMEar(t) (4) The Fourier coefficients (CHi, Cvi) and phases ($//,-, <j>Vi)
where MME is the structural mass of the test pipe and the effect have been determined by a direct Fourier transformation and
of the entrapped water. MME is known and the last term of a least-squares-fit technique.
Eq. (4) is subtracted from the measured force before the anal-
ysis is performed. Extreme Force Analysis. The extreme force coefficients
were determined by normalizing the extreme forces in each
FH(t) -MMEar{t) = i PDCDU(t) I U(t) I half wave by l/2pDU2w. The missing Froude-Krylov term -7r/4
pL^-ait) was added to the measured in-line force before the
analysis was performed in order to determine coefficients valid
+ ^pD2Caar(t) (5) for the prototype situation.
The extreme force coefficients were determined as
Therefore, Eq. (5) was used in the analysis. The output from
the analysis was CD and C, and the inertia coefficient was CH,mm = F(kt)mn/ ~ pDU2w (12)
found as
CM = 1 + C. (6) Cv,m^=FL(t)m3X/-PDUl (13)

Fourier Analysis The extreme coefficients were divided in two groups, one
for forces in zero-upcrossing half-periods and one for forces
Previous work (e.g., references [1-5]) has demonstrated that in zero downcrossing half-periods. These coefficients were sub-
the commonly used Morison-type formulations yield force pre- sequently statistically analyzed, and mean values and standard
dictions which often give a poor description of the measured deviations determined.
forces.
Therefore, a mathematical approach was employed in order
to obtain more accurate predictions in the case of regular Test Results
waves, namely, the Fourier decomposition method. The basis The results of the model tests are the three sets of force
for this method is that any quantity that has a periodic variation coefficients:
with a period 7, can be reproduced by superposition of a
number of sine waves with periods equal to 7 and smaller, so 1 Drag, inertia and lift coefficients (CD, CM, CL)
that the /th sine wave or superharmonic has a period 7} = 7 / 2 Fourier coefficients and associated phase angles (C H , 4>Hn,
/, / = 1, 2, 3 . . . N. For the present study, it was assumed Cy, <l>yn)
that 9 harmonics and a mean value would result in acceptable 3 Extreme force coefficients (CHiVmx, CV.max)
accuracy of the predicted forces.
The general expression of the periodic quantity F(t) is, thus, which, together with the associated analytical expressions, yield
given by force predictions of very different accuracy.
N The drag, inertia and lift coefficients in combination with
F(t) = O 0 + 2 J ( a ' cos
iwt+bj sin icat) (7) the Morison-type formulas for in-line and lift force yield rea-
I sonable force predictions for the in-line force, whereas the lift
force in most cases is poorly described for both regular waves
or with and without steady current superimposed.
N
The Fourier coefficients in combination with the nine-term
F(t) = C0+^Q cos Hut-fr) (8)
force expression give in all cases very accurate description of
l
the hydrodynamic forces in regular waves and combined cur-
where C, = (a2 + b2)l/2, tan (/ </>,) = bi/ah w = 2-w/T, and rent and regular waves.
a, and 6, are the Fourier coefficients to be determined. The extreme force coefficients give accurate estimates of the
In order to determine nondimensional force coefficients from maximum in-line and lift force in each half-wave period. How-
the Fourier analysis, the measured forces were normalized by ever, no information regarding time variation or phase angle
1/lpDU2,, Uy, being the wave velocity amplitude. This nor- between the in-line and lift force is obtained.
malization would result in extremely large coefficients for the For further discussion on force predictions, see reference
horizontal force at small KC number, where the force is totally [10].

234 / Vol. 114, NOVEMBER 1992 Transactions of the ASME

Downloaded 15 Jan 2013 to 223.27.129.246. Redistribution subject to ASME license or copyright; see http://www.asme.org/terms/Terms_Use.cfm
 1
t i i ' 1 1 . , 1 ' 1 ' 1
' ' CO
UGHNE35
r
,\ FJOU H

iA -
- "
,\ \\\r^^ ^ -

J*~^
_ ~~~~- r^
VI
II

KC

"1 1 1 11 * 1 1 1 1 1 1 1 1
=
1 I 1

PIPE SURFACE ROUGHNESS - A '_

FINE
. . . ,ET.l
... PoieHT,i r u m wuje
_._ou. -
BOTTOM nOUOHNCSS " k M KC 70 .
FINE

- - V ... - - >^ -
^ " ^ ^ ^ S T " . 1
- \\ "
u u
^ ~ ' - - "

' 1 . 1 , 1 i , i , t , i , T" - 1 1 1 "

- -
(a) (*) _
Fig. 6 Force coefficients CD, Cu, and CL for regular waves(a) effect - HZ
of pipe surface roughness, (b) effect of seabed roughness
Fig. 7 Drag, inertia and lift coefficients versus current ratio, a, for
regular waves and current (three pipe surface roughnesses are
shown: fine ( ), medium ( ) and rough (--))
Force Coefficients for Drag, Inertia, and Lift
Regular Waves. The results of all regular wave tests with-
out superimposed steady current are shown in Fig. 6. The force a-values the larger the KC number. The inertia coefficient
coefficients are plotted against the wave parameter, KC. Two decreases with increasing a-value. The lift coefficient decreases
sets of coefficients are presented. In Fig. 6(a), the coefficients with increasing a and the effect of a steady current can be
are shown for the three different values of pipe surface rough- compared to the effect of increasing the KC number in case
ness, and in Fig. 6(6), the results for three different values of pure wave action. This effect of the steady current is most
of the seabed roughness are shown. The full lines in the two pronounced for small KC numbers.
sets of graphs represent the same test series. Effect of Pipe Surface Roughness. The pipe surface con-
As indicated on the graphs, the variation with the wave ditions tested vary from relatively smooth conditions to ex-
parameter (KC) is very similar for all values of the pipe surface tremely rough. Practical design situations will normally fall
and the seabed roughness. The drag coefficient CD increases within the conditions investigated. Figure 6(a) shows the effect
for increasing values of KC in the interval 2<KC<10-15. of pipe surface roughness in pure wave action and Fig. 7 shows
Around KC = 10-15, the drag coefficient attains its maximum the effect in combined wave and current action. It can be
value for thereafter to decrease with increasing KC-number. observed that the drag and inertia coefficients increase sub-
The inertia coefficient CM attains values close to its theoretical stantially with the surface roughness, whereas a slight decrease
potential flow value of 3.29 for small KC-numbers. For larger can be seen for the lift coefficient. Expressed in hydrodynamic
KC-numbers, the inertia coefficient attains larger values, e.g., forces, the in-line force increases, whereas the lift force remains
for KC = 160, CM is larger than 5. The lift coefficient CL unchanged or decreases slightly. The observed influence in the
exhibits a variation with the KC number, which is fairly simple. pure wave case becomes more significant when a steady current
For small KC-numbers, CL approaches the theoretical poten- is superimposed. Both the increase in drag force and the de-
tial flow value of 4.49. For increasing KC numbers, CL rapidly crease in lift force are relatively stronger.
decreases and approaches 1.0 for KC = 160.
Effect of Seabed Roughness. Under oscillatory flow con-
Regular Waves and Current. A representative selection of ditions the seabed roughness gives rise to the formation of a
the results from combined regular wave and steady current thin unsteady boundary layer. The effect of this boundary
tests is shown on Fig. 7 as plots of the force coefficients against layer on the hydrodynamic forces has been investigated by
the current ratio, a. Each graph presents test results for one conducting tests in which the seabed was provided with rough-
KC-value and three different pipe surface roughnesses. ness having three different values from relatively smooth to
The presence of a steady current leads in all cases to reduction rough conditions. Figure 6(6) shows the effect of the seabed
of the force coefficients. The drag coefficient approaches the roughness on the force coefficients. Increasing roughness leads
value found in steady current for increasing a-values. It is to reduced drag and lift coefficients, whereas the inertia coef-
characteristic that the steady current value is found for lower ficient is insignificantly influenced.

Journal of Offshore Mechanics and Arctic Engineering NOVEMBER 1992, Vol. 114/235

Downloaded 15 Jan 2013 to 223.27.129.246. Redistribution subject to ASME license or copyright; see http://www.asme.org/terms/Terms_Use.cfm
 ,.l"in <
s-s^

Fig. 8 Fourier coefficients and phases for the in-line drag force (regular Fig. 10 Fourier coefficients and phases for the in-line drag force (reg-
waves, pipe and seabed roughness: fine) ular waves and current, KC = 40, pipe and seabed roughness: fine)

Fig. 9 Fourier coefficients and phases for the lift force (regular waves, Fig. 11 Fourier coefficients and phases for the lift force (regular waves
pipe and seabed roughness: fine) and current, KC = 40, pipe and seabed roughness: fine)

Fourier Coefficients ficients and phases for the drag and lift force is shown in Figs.
10 and 11, respectively. The coefficients are no longer directly
Regular Waves. The Fourier coefficients and phases for comparable with the traditional coefficients CD and CL. Both
the drag force and the lift force are shown in Figs. 8 and 9, for the drag and the lift force, a significant input is provided
respectively. Only results from the basic test conditions cor- by all coefficients, even as well as odd, and their associated
responding to the smallest pipe and smallest seabed roughness phase angles. Coefficients of higher order than five are small
are shown. The data for increased seabed roughness exhibit enough to be neglected in all practical applications. The com-
the same variation as shown in Figs. 8 and 9. Considerations plex formulation makes it impossible to clearly identify and
of symmetry disclose that all even coefficients (Co, C2, . . C2) describe the influence of the governing parameters based on
theoretically equal zero for the drag force and all odd coef- simple evaluations of the coefficients.
ficients (Cu C3, . . C2+i) theoretically equal zero for the lift
force. Extreme Force Coefficients
Comparing the dominating Fourier coefficients with the cor-
responding drag and lift coefficient, it can be seen that the Regular Waves. The extreme force coefficients found by
coefficients exhibit a similar variation with the KC number. normalizing the maximum forces with the factor \/2pDU2w are
Taking into account the similarities between the simple Mor- shown in Fig. 12 for the three pipe surface roughnesses. The
ison-type formulation and the full Fourier expansions, this is variation with the KC number of CHtmax is similar to the Co-
not surprising. On the other hand, it is also clear that the phase variation for the large KC numbers. For small KC numbers,
angle and the higher order coefficients provide significant con- where the inertia force is dominant, the limiting value equals
tributions to the force predictions of the Fourier expansion. 32.5/KC (ir2CM/KC). This value is found assuming CD = 0
and CM =3.29 and the data thus confirm the theoretical value
Regular Waves and Current. An example of Fourier coef- of the inertia coefficient. Cv,mSx has a variation with KC quite

236 / Vol. 114, NOVEMBER 1992 Transactions of the ASME

Downloaded 15 Jan 2013 to 223.27.129.246. Redistribution subject to ASME license or copyright; see http://www.asme.org/terms/Terms_Use.cfm
 1 1 Table 3 Test matrix for irregular waves (JONSWAP spec-
fxCHMAX ! PI
1 < 1 '
P E SURFACE ROUGHNESS
' I '

I X trum has been applied generally, PM spectrum additionally

o u " . FINE
- T. x
A li* . MEDIUM
for KC = 20, 30 and 50)
W X k * , ROUGH
Current ratio
tv KC
no. 0.0 0.16 0.32 0.48 0.64 0.80 1.20
J X
X 10 ,<) 1 1 1 1 1
KC
6 X
X 15 1 1 1 1 1 1
CT * X
6 ca OA oa 20 2 1 1 1 1
30 2 1 1 1
KC 50 2 1 1
I 1 i 1 i 1 1 ,
70 1

i i j t i i | ) ,,
... f.. r
r
J
X C V MAX , PIPE SURFACE ROUGHNESS
" 1 " indicates that one or more tests have been performed
X
O k* . FINE -

a /\
% X *** ROUGH
- %9
C6 REGULAR WAVES
*<* IRREGULAR WAVES
x2 6
X
X
* St _
Q>
U 8
<* * ".
KC
i I i i 1 1 i i 1 1 >

Fig. 12 Extreme force coefficients versus Keulegan-Carpenter number

for regular waves

J i I I I i l_
0 20 -0 60 80 100 120 IfcO 160

-iiiir

QO 3 29 POTENTIAL FLOW VALUE

ZLIJO ufJCPOSSirJG
ZERO DOWMCROS5

REGULAR WAVES
IRREGULAR WAVES

20 to 60 80 100 120 I&0 160

Fig. 13 Extreme force coefficients versus current ratio for KC-numbers

of 40 and 70 1 1 1 I 1 ' 1
_CL

REGULAR WAVES
\ FLOW VALUE
IRREGULAR WAVES

similar to CL, but CV,max is larger than CL for intermediate t

KC-numbers. The large lift forces are associated with high \
0
flow velocities near the pipe induced by the reversal of the
- \,
wake [6].
Regular Waves and Current. The superposition of a steady
current leads in general to extreme forces that are larger in KC -
half-periods where wave and current velocities add than in the 0
, 1
20
. i
i.0
, !
60
, 1
80
, 1
100
1 1 1 1 1 1
120 1A0 160
half-periods where the two velocities counteract each other.
Fig. 14 CD, CM, and CL versus KC for irregular wave flow
However, as can be seen from Fig. 13 there is no significant
difference in the force levels in two succeeding half-waves for
current ratios below approximately 0.35. Force expressions
based on velocity-squared terms only would in fact predict Irregular Waves
force ratios in succeeding half-waves of the order of (1 + A number of irregular wave tests with and without steady
0.35) 2 /(l-0.35) 2 = 4.3 for a = 0.35. It is the dominant in- currents was conducted according to the test matrix shown in
fluence of the wake reversal that leads to relatively high-force Table 3.
levels also when the free stream wave velocity opposes the The forces measured were analyzed with the LSF method
steady current. Increasing a-values reduce the effect of the based on Eqs. (1) and (2) to yield CD, CM, and CL applicable
wake reversal. However, the ratio of the peak forces in two for the full test duration. Figure 14 shows the results compared
succeeding half-waves never reaches the level predicted by the to the regular wave test data. For the irregular waves KC was
free stream velocities squared. It is obvious for very large values based on the significant velocity and the peak period of the
of a, both peak forces should become equal. velocity spectrum. The variation of the force coefficients with

Journal of Offshore Mechanics and Arctic Engineering NOVEMBER 1992, Vol. 114 / 237

Downloaded 15 Jan 2013 to 223.27.129.246. Redistribution subject to ASME license or copyright; see http://www.asme.org/terms/Terms_Use.cfm
 " ' KC-78
cD KC 10 cD C > 40 IRREGULAR" UAVC3 o . KC-ao
SEOULAR nffULAfl
- - KC 5 0
WAVE 3
c > ao
KC . JO KC .70

one ~ta

w i v n
IKC
t KC - a o
1B ' TJ.T" KC - 7 0

0=
" ' 0
b o
^x
\ ? ^-i-.W_ ^ >^
>==-- "
* "r-

r - a , iii. Msi
E^JL"
8 V - - "
,,
77~^r*?-

C M
1IOUIAH
W A V 3

KC JO

0 C 10 O KC fc3
IRREOULAR UAVE3
IflRHJUUD **C .MS
D . KC-2B
w i t f l l

ss:r.. 0
-^
* - -
-e.--i,u
^ -
x"
St

1l ..
c, "- v ? * - - -
KC BO ",... 'C: . ;
. KC 0
- - * e 30 - KC t 70 .
O KC ~%o OKC " * 3
\ IRR10ULAR X MC " I S XKC - 7 0
\ W A * 4 3 4 KC
o c
i \ -JO .

\V
"\V^X
Fig. 16 Extreme force coefficients versus local KC-numbers. Results
from irregular wave tests having global KC-numbers of 20 and 70. Solid
line: result from regular wave tests.
5==Z- tV^o
^XSXZ=3Z^-
.

a. Si-
a >
fe- have been applied. The following data have been compared
with the results of the present study:
Fig. 15 Co, CM, and CL versus a. Irregular waves and steady current
1 Laboratory tests carried out at the Naval Postgraduate
School, Monterey, California, reported by Sarpkaya and Ra-
the KC number is quite similar for the two test conditions. jabi, 1980 [5].
The drag coefficients are generally smaller for irregular waves, 2 Laboratory tests carried out at the Norwegian Hydrody-
whereas the lift coefficient attains values almost identical to namic Laboratories (NHL) as part of a joint industry project,
those for the regular wave tests. reported 1985-1986 [7].
Data for combined wave and current flow are shown in Fig. 3 Field experiments carried out off the coast of Hawaii; partly
15 and also here the irregular wave results display a variation sponsored by A.G.A. Reported by Grace, 1979, 1981, [2, 8].
similar to that of the regular tests.
The Fourier decomposition cannot be applied directly to the Despite some differences in analysis techniques and actual
irregular wave forces, as a very large number of harmonics values of testing parameters, these three references form a good
would be required in analogy with spectral analysis of irregular basis for comparison.
waves.
Extreme force coefficients were found for each half-wave
Comparison for CD, CM, and CL with Data from Sarpkaya
by normalizing the maximum forces with the value l/2pDU2M,
and Rajabi [5]. Based on the Morison equation, Sarpkaya
where 4, is the maximum wave-induced velocity in the relevant
and Rajabi have derived CD and CM values as so-called Fourier
half-wave period. Local KC numbers were defined based on
averages
the UM and the associated half-periods, and the extreme force
coefficients were plotted versus these local KC numbers. Figure
CD= (Vi/2PDU2m)- F(0) cose dd (14)
16 shows the extreme force coefficients for two irregular tests,
one having global KC number of 20 and the other 70. The ATT
scatter is quite significant, but the average trend follows that
of the regular wave data shown in solid line.
KC
{ o F(B) sine dd (15)

where Um is the velocity amplitude for an oscillating flow

Comparisons With Other Data represented by U = Um- sin0, 6 = 2-Kt/T, 71s the oscillation
Experiments to determine the hydrodynamic forces on pipe- period, and Fm{6) is the measured in-line force. This analysis
lines have also been carried out by other researchers. These method yields results very similar to a least-squares-fit analysis
experiments have investigated the influence of one or more of for a perfect sinusoidal flow, as demonstrated by Sarpkaya
the governing parameters, and various data analysis methods [9].

2 3 8 / V o l . 114, NOVEMBER 1992 Transactions of the ASME

Downloaded 15 Jan 2013 to 223.27.129.246. Redistribution subject to ASME license or copyright; see http://www.asme.org/terms/Terms_Use.cfm
 o L_i l i l i i i i i i i i . i i i
0 20 i0 60 80 100 120 1 /.0 160
i 1 i | 1 | 1 1 1 1 : 1 1 1 r,
7 _ CM : l ,ca Q_

Fig. 19 Fourier coefficients and phases for the lift forceregular waves

The smooth pipe data from Sarpkaya and Rajabi show

slightly smaller Co-values in accordance with the smaller k/
D-value in their tests. The inertia coefficient appears to be
slightly larger than the present data, but the overall agreement
between the two sets of data is good, and a C^-value close to
3.29 is reached at low KC numbers.
For the rough pipe case, reference [5] yields somewhat larger
force coefficients than the present ones for corresponding rel-
n
3
I ' I
20
' I
40
i I
60
l i
60
i I
100
. I
120
'
: L0
i
16C
ative pipe roughness.
Fig. 17 Data for CD and CM from reference [5] and from present study
On Fig. 17, the data from the recent PIPESTAB study are
added [11]. The tests have been performed with the same setup
as the tests in reference [7]. The coefficients, CD and CM, from
reference [11] are in good agreement with the other studies.

Comparison for Fourier Coefficients and Phases With Data

from NHL [7]. In the NHL study the measured in-line and
lift forces in regular wave motion with and without a steady
current have been decomposed into Fourier series using a tech-
nique similar to the one used in the present study. The measured
in-line force, however, has been treated somewhat differently
by NHL as the Fourier amplitudes and phases include drag
forces as well as inertia forces.
The data from NHL [7] have been re-plotted in Figs. 18 and
19 together with the data from this study. The figures show
the Fourier coefficients and phases for pure wave motion. The
test conditions for the two sets of data have been quite similar.
The NHL data cover experiments with two pipe diameters of
0.2 m and 0.5 m.
The agreement between the two sets of in-line force data is
excellent, Fig. 18. Deviations appearing at small KC numbers
in the first harmonic amplitudes and phases are due to the
inertia force being included in the NHL data. This part of the
force is dominating at low KC numbers, and it varies with the
Fig. 18 Fourier coefficients and phases for the in-line forceregular
waves
oscillation period, i.e., the period of the first harmonic. There-
fore, an increase in the first harmonic amplitude appears and
a phase lag is introduced. For the larger KC numbers, the data
For the lift force, Sarpkaya and Rajabi have not applied a correlate well, and for the higher harmonics the excellent agree-
least-squares-fit type of analysis, but they have determined' ment covers the full parameter range. The minor deviations
maximum lift coefficients comparable to the extreme force may well be explained by differences in pipe surface roughness.
analysis. NHL used a "smooth" pipe surface presumably having a
The data for CD and CM from reference [5] have been plotted k/D value lower than 10 -3 used for the present data.
together with the present data on Fig. 17. From Fig. 17, it is The Fourier amplitudes and phases for the lift forces are
seen that the drag and inertia coefficients from the two sets generally in good agreement for all orders of the harmonics
of data exhibit the same variation with the KC number and as shown in Fig. 19. There is, however, a tendency toward a
that increased pipe surface roughness leads to increased force larger constant lift force coefficient, Cot by the NHL data as
coefficients. compared to the present ones. This observation is in agreement

Journal of Offshore Mechanics and Arctic Engineering NOVEMBER 1992, Vol. 114/239

Downloaded 15 Jan 2013 to 223.27.129.246. Redistribution subject to ASME license or copyright; see http://www.asme.org/terms/Terms_Use.cfm
 KC-7B
OMI (AGA 1986) IRREGULAR WAVES
0 . KC-2B

< 15" i
0 mma GRACE I i98i
\
0o
0

* *,'*"
.*.i*.v,iit . * ,
o :I?V
!v
0
' " J * .*
i>

HI
/

0
iir 1
i^T - .". - .i -\" . -.
VARIATION OF MAXIMUM VERTICAL FORCE COEFFICIENT
WITH ADAPTED PERIOD PARAMETER FOR oC = 15 '
40 0 00.0

Fig. 21 Coefficients for the extreme horizontal forces, CH,

O H M A O A 19Bfl> IRREGULAR WAVES

1 r
l*:*SS-l GRACE ( 1 9 8 1 )

VARIATION OF MAXIMUM HORIZONTAL FORCE COEFFICIENT

WITH ADAPTED PERIOD PARAMETER FOR oC - 1 5 *

Fig. 20 Data for C Hmax and Cymax from reference [8] { l

with the assumption that the NHL test pipes have had a rough-
ness ratio less than 1CT3. Fig. 22 Coefficients for the extreme lift forces, Ci/ m
The excellent agreement between the two data sets including
the higher harmonics indicates that the actual force time series
from the two studies have been very similar for identical test 1 1 1 1 I 1 ' 1 ' 1 '
CH MAX GRACE
conditions, a presumption that has been further supported by DHI
visual comparisons of measured force time series. \
\
Comparison for Extreme Force Coefficients With Data From - \\ -
Grace [8]. From time series of measured in-line and lift forces
and corresponding time series of water particle velocities and
- V "
accelerations, the peak forces occurring in each wave cycle
have been normalized with l/2pDU2w by Grace, where U is
the maximum particle velocity measured during the same wave - --. -~ -
cycle. This normalization procedure is similar to the one used
in the model tests. However, Grace looked at a full-wave cycle,
where a half-wave by half-wave analysis has been applied in - KC
the present study. . i . 1 , 1 , 1 , 1 , 1
The results by Grace were plotted in reference [8] in terms
Fig. 23 Extreme in-line force coefficients. Field test data and regular
of peak force coefficients as function of an adapted period
wave data from present study.
parameter, \p = U^UD, where Uw is the maximum acceler-
ation of the water particles. For regular waves this parameter
is equal to KC/2ir and this relation has been applied for the tests yield somewhat larger vertical force coefficients, espe-
comparison although the relation is not fully valid for irregular cially for larger KC numbers, Fig. 22. This may be partly due
waves. Figure 20 below shows the data from reference [8] for to the fact that a small gap was present between the pipe and
the case with an angle of 15 deg between the pipe and the wave its base structure in the field experiments. Such a gap reduces
fronts. These data have been used for the comparison. the maximum lift forces, but does not change horizontal forces.
Figures 21 and 22 show the comparison between the results Furthermore, the half-wave analysis applied in the present
of two irregular wave tests with KC equal 20 and 70, and the study will lead to a larger number of extreme lift coefficients
field test data by Grace, shown as the hatched area. Note that with high values. This is because the maximum velocity of each
the Grace data have been divided by cos 2 15 deg in order to half-wave is used for the normalization of the measured force,
make the results comparable with the present model test results. and large lift forces may occur in half-periods with small ve-
The first observation from the two figures is that the scatter locities when the preceding half-periods are associated with
is similar for the two sets of data, and that the general trend large velocities.
with the KC number is identical. The extreme horizontal force The mean values of the coefficients from the field tests are
coefficients correlate excellently, Fig. 21, whereas the model shown in Figs. 23 and 24, together with the data for regular

2 4 0 / V o l . 114, NOVEMBER 1992 Transactions of the ASME

Downloaded 15 Jan 2013 to 223.27.129.246. Redistribution subject to ASME license or copyright; see http://www.asme.org/terms/Terms_Use.cfm
 1
/ u
C
1 i > r 1 ' 1 ' 1 ' ness ratio. The parameter ranges tested cover most situations
V MAX GRACE
DHI met in practical design.
6 0
" N jIiP&3 SARPKAYA The results have been compared to three other studies and
excellent agreement was found, not only in relation to force
5 0 - coefficients, but also with respect to time variation of the
forces.
\
U0
- \ Acknowledgments
^
3 0
*".
-^ The findings presented in this paper are based on work
carried out by the Danish Hydraulic Institute for the American
2 0
'^v^-~^- Gas Association, 1983-1986. The authors wish to express their
"^""41
appreciation to the American Gas Association for permission
1 0
to publish this paper. The authors would also like to thank
the members of A.G.A.'s project ad-hoc committee for their
KC
cooperation, and Mr. R. F. Schollhammer of the A.G.A. for
O , i 1 , 1 t , i . i .
his support.
Fig. 24 Extreme lift force coefficients. Field test data and regular wave
data from reference [5] and from present study.
References
1 Bryndum, M. B., Jacobsen, V., and Brand, L. P., "Hydrodynamic Forces
from Wave and Current Loads on Marine Pipelines," Proceedings of the Fif-
waves from the present study. Again, the agreement for the teenth Annual Offshore Technology Conference, Houston, Texas, OTC 4454,
horizontal force is good, and larger values for the lift force May 1983, pp. 95-102.
2 Grace, R. A., and Zee, O. T. Y., "Wave Forces on Rigid Pipes Using
are found in the model tests. The maximum lift force coef- Ocean Test Data," Journal of the Waterway, Port, Coastal and Ocean Division,
ficients from reference [5]regular wavesare also included ASCE, Vol. 107, No. WW2, 1981, pp. 71-92.
in Fig. 24, and these data compare very well with the DHI 3 Littlejohns.P. S. G., and Spencer, J., "Wave Forces on Pipelines," Report
data. Nos. EX 1056 and EX 1059, Hydraulic Research Station, Wallingford, England,
1982.
4 Yamamoto, T., Nath, J. H., andSlotta, L. S., "Wave Forces on Cylinders
Near Plane Boundary," Journal of the Waterways, Harbours and Coastal En-
Summary and Conclusion gineering Division, ASCE, Vol. 100, No. WW4, 1974, pp. 343-360.
5 Sarpkaya, T., and Rajabi, F., "Hydrodynamic Drag on Bottom-Mounted
The hydrodynamic forces on submarine pipelines resting on Smooth and Rough Cylinders in Periodic Flow," Proceedings of the Eleventh
the seabed have been determined through model tests. The Annual Offshore Technology Conference, Houston, Texas, OTC 3761, May
following environmental conditions have been investi- 1980, pp. 219-226.
6 Jacobsen, V., Bryndum, M. B., andFreds0e. J., "Determination of Flow
gated: steady current, regular waves, combined steady cur- Kinematics Close to Marine Pipelines and Their Use in Stability Calculations,"
rent and regular waves, irregular waves, and combined irregular Proceedings of the 16th Annual Offshore Technology Conference, Houston,
waves and steady current. Texas, Paper OTC 4833, May 1984.
7 Norwegian Hydrodynamic Laboratories, "Design of Pipelines to Resist
The tests have been analyzed using three different methods, Ocean Forces," Final Report on Joint Industry R&D Program, 1985.
i.e., a least-squares-fit analysis in combination with Morison 8 Grace, R. A., " A Slanted Look at Ocean Wave Forces on Pipes," Report
type of equations, Fourier analysis and maximum force anal- to the American Gas Association, Aug. 1979.
ysis. The force coefficients presented together with the ana- 9 Sarpkaya, T., "Forces on Cylinders Near a Plane Boundary in Sinusoidally
lytical force expressions constitute a very comprehensive and Oscillatory Fluid," ASME Journal of Fluids Engineering, Vol. 98, 1976, pp.
499-505.
consistent data base applicable for calculating hydrodynamic 10 Jacobsen, V., Bryndum, M. B., and Tsahalis, D. T., "Prediction of Ir-
forces on submarine pipelines. The influence of the main non- regular Wave Forces on Submarine Pipelines," Seventh Offshore Mechanics
dimensional governing parameters on the hydrodynamic force and Arctic Engineering Conference, Houston, Texas, Feb. 1988.
has been identified and quantified. The parameters investigated 11 Verley, R. L. P., Lambrakos, K. F., and Reed, K., "Prediction of Hy-
drodynamic Forces on Seabed Pipelines," Proceedings of 19th Annual Offshore
were the Keulegan-Carpenter number, the current to wave Technology Conference, Houston, Texas, Paper No. OTC 5503, Houston, Texas,
ratio, the pipe surface roughness ratio, and the seabed rough- May 1987.

Journal of Offshore Mechanics and Arctic Engineering NOVEMBER 1992, Vol. 114 / 241

Downloaded 15 Jan 2013 to 223.27.129.246. Redistribution subject to ASME license or copyright; see http://www.asme.org/terms/Terms_Use.cfm