Академический Документы
Профессиональный Документы
Культура Документы
Journal of Offshore Mechanics and Arctic Engineering MAY 2007, Vol. 129 / 107
Copyright © 2007 by ASME
冉 冊 冉 冊
D Ph
Dt w
+
D
Dt
f
t
D
+ 共gz兲 = 0
Dt
共7a兲
ⵜ ⌽=0
2
where
Ph ⌽ 1 共2兲
+ + gz + 兩ⵜ⌽兩2 = F1共t兲 D
w t 2 = +q·ⵜ 共7b兲
Dt t
where F1共t兲 is constant in space and for this purpose assumed as
zero. In a Cartesian coordinate system, Eqs. 共2兲 can be written as and
2⌽ 2⌽ 2⌽
+ +
x2 y 2 z2
=0 共3a兲 q= f 冉 f
,
x z
冊 共7c兲
Ph ⌽
+
w t
+ gz +
1
2
冋冉 冊 冉 冊 冉 冊 册
⌽
x
2
+
⌽
y
2
+
⌽
z
2
= 0, ∀ z = a
Thus for isentropic 共sea state兲 inviscid fluid flow, Eq. 共7a兲 now
takes the form
1 Ph 2 f z
共3b兲 + 2 +g =0 共8兲
w t t t
Since ⌽ = ⌽共x , y , z , t兲 satisfies Eq. 共3a兲, we introduce the variable
separable solution Now, at z = a, we invoke the kinematics boundary condition,
namely
⌽ = 共x,y,兲f共z,t兲
f
which allows us to rewrite Eq. 共3a兲 as = 共9兲
t z
But at z = a, Ph = Pa 共atmospheric pressure兲, which allows us to
rewrite Eq. 共8兲 as
2 f f
2 + g =0
t z
or as
2 f f
=−g 共10兲
t2 z
Now, the governing equation of motion in domain 共⍀2兲 as recalled
from Eq. 共1兲 admits the form
EIWxxxx + 共 + f 兲Wtt + 2u f Wtx + f u2Wxx
PS
+ KbW − S␦共R − ␦/2兲
x
= Ph共2R − ␦兲 ∀ x 苸 ⍀2 共internal兲 共11兲
where, ␦ is the level of sediment coverage of the pipeline at a
given time. Thus
ẑ = z + h + R
and
冏 冉
Ph = − wzg + w
⌽
t
冊冏 z=−h
共12兲
Journal of Offshore Mechanics and Arctic Engineering MAY 2007, Vol. 129 / 109
f̃共z,s兲 = −
sW̃共x,s兲
cosh kz 共20兲
i.e. k sinh kh
ⵜ2共x,y兲
共x,y兲
= − k2 ;
2 f共z,t兲
z2
冒 f共z,t兲 = k2 共13b兲
Now, the reduced velocity potential 共x , y兲 is defined from
共x,0兲
= UW
where k is a constant 共wave number兲 to be determined from the
2 x
fluid kinematics boundary conditions. Hence
冕
The first of Eqs. 共13兲 is the Helmholtz equation, namely L
ⵜ +k =0
2 2
共14兲 共x,0兲 = UW dx = UWL 共21兲
0
where 共x , y兲 is the reduced fluid velocity potential.
where Uw= uniform water velocity at infinity 共for the sea state
The second of Eqs. 共13兲 gives
earlier defined兲.
2 f共z,t兲 We now introduce the Laplace transform earlier used as namely
冕 冕
− k2 f共z,t兲 = 0 共15兲 ⬁ ␥+i⬁
z2 1
共•˜ 兲 = 共•兲e−st dt, 共•兲 = 共•˜ 兲est ds 共22兲
subject to the boundary conditions 0
2i ␥−i⬁
f共z,t兲 W
− S␦共R − ␦/2兲
dPS共s,x兲
dx
= 共2R − ␦兲w − s⌽̃共x,s,z兲 +
gh
s
冋 册
= at z = − h, in ⍀2 共16b兲 共23兲
z t
If we consider no external excitation in ⍀3 at this stage, we can
Equation 共16a兲 represents the free surface 共no surface tension兲 sea
write
state boundary condition and Eq. 共16b兲 represents the fluid–
structure–seabed dynamic interaction boundary condition. EIW̃xxxx共x,s兲 + 共 + f 兲关s2W̃共x,s兲兴 + 2u f 关sW̃共x,s兲兴 + f u2W̃xx共x,s兲
The general solution of Eq. 共15兲 in the Laplace transform plane
admits the form dPS共s,x兲
+ KbW̃共x,s兲 − S␦共R − ␦/2兲
x
冋 册
f̃共z,s兲 = A共s兲ekz + B共s兲e−kz 共17兲
gh
while the Laplace transform of the boundary conditions 共16a兲 and = 共2R − ␦兲w − s⌽̃F共x,s,z兲 + 共24兲
s
共16b兲 gives
wherein for motion through the domain 共⍀3兲
共z,s兲 1
df
= 关− s2 f共s,z兲 − sf共0,z兲 − ḟ共0,z兲兴, z = 0 苸 ⍀1 PS = PS共x,t兲
dz g
We next introduce Fourier finite sine transform namely
冕
L
共z,s兲
df nx
= sW̃共x,s兲 − W共0兲, z = − h 苸 ⍀1 共external兲 共18兲 IS关兴 = 关•¯ 兴 = 关•兴sin dx
dz 0
L
再 冎
On substitution of the zero initial conditions, we find ⬁
nx
兺
2
共z,s兲 关•兴 = 关•¯ 兴sin 共25兲
df s2 L n=1 L
= − f̃共s,z兲
dz g Thus, Eq. 共24兲 can be written as
冏 冏
共z,s兲
df
dz z=−h
= sW̃共x,s兲 共19兲
EIIs共W̃xxxx兲 + 共 + f 兲s2Is关W̃共x,s兲兴 + 2u f sIs关W̃x共x,s兲兴
+ f u2Is关W̃xx共x,s兲兴 + KbIs关W̃共x,s兲兴
Consequently, we obtain
− S␦共R − ␦/2兲␦Is 冋 dPS共s,x兲
册
s
k
冉 冊
W̃共x,s兲e2kh 1 +
s
gk s
2
冋
= 共2R − ␦兲w − s⌽̃F共n,s,z兲 +
dx
gh F
册
冉 冊
A共s兲 = − 2 + W̃共x,s兲ekh 1 共26兲
s k s
2 sinh kh + cosh kh
gk where
冋 册
efficient; Us⫽seepage horizontal velocity component; C p= poros-
dPS共x,s兲 ity coefficient or index; and s⫽bed sliding friction coefficient.
+ KbW̃F共n,s兲 − S␦共R − ␦/2兲IS For this case
dx
冋
= 共2R − ␦兲W s⌽̃F兩共n,s,z兲兩z=−h +
gh F
s
1 册 共30兲 Is 冉 冊
dP̃s
dx
,=−
sc pU s F
sKs
1
If we invoke the decomposition of ⌽, introduced earlier, we can and Eq. 共34兲 simplifies to
write from Eqs. 共20兲 and 共21兲
s cosh kz W̃ 共n,s兲 =
F
冉
共2R − ␦兲 wgh −
s2␦c pUs F
2Ks
1 冊 共37兲
⌽̃ 共n,s,z兲 = 共n兲f̃共z,s兲 = − UWL W̃共x,s兲1F
F F
关共 + f 兲 − 共2R − ␦兲w兴s共s2 + 2兲
k sinh kh
So that we can write
= ˆ sW̃F共n,s兲 共31兲
⌸
And using Eq. 共31兲, we write Eq. 共30兲 in the form of W̃F共n,s兲 = 共38兲
s共s2 + 2兲
n 4 4 F 2n
2 2 where
4 W̃ 共n,s兲 + 共 + f 兲s W̃ 共n,s兲 − f u
冉 冊
EI 2 F
W̃F共n,s兲
L L2 s2␦c pUs F
冋 册
共2R − ␦兲 wgh − 1
dPS共x,s兲 2Ks
+ KbW̃F共n,s兲 − S␦共R − ␦/2兲IS ⌸= 共39兲
dx 共 + f 兲 − 共2R − ␦兲w
共2R − ␦兲Wgh1F This gives us the Laplace inversion of W̃F共n , s兲 as
= 共2R − ␦兲Ws2W̃F共n,s兲 + 共32兲
s ⌸
WF共n,t兲 = 共1 − cos t兲 共40兲
where 2
U WL and
ˆ = − cosh kz ⬁
k sinh kh ⌸ nx
兺
2
W共x,t兲 = 共1 − cos t兲sin
and L n=1 2 L
⬁
⌸̂ 共1 − cos ˆ t兲 2nx
兺
U WL
兩 = ˆ 兩z=−h = − coth kh 共33兲 = sin 共41兲
k L n=1 nˆ 2
L
and using Eq. 共31兲, we cast Eq. 共30兲 in the form of where
Journal of Offshore Mechanics and Arctic Engineering MAY 2007, Vol. 129 / 111
冋冉 冊 册
x t Ks ⬁
␣ˆ 1 + ␣ˆ 2 ␣ˆ 3US 2nx
兺
1
where W共x,t兲 = 共1 − cos ˆ t兲 − sin ˆ t sin
L n=1 nˆ 2
nˆ L
共Us − Ub兲 = sediment-seabed-relative velocity
共47兲
Now
where
dP̃s共x,s兲 cp
= − s关sŨs共s兲 − U共0兲兴 − 共 U s − U b兲 s ␣ˆ n = ␣n/1F
dx sKs
in the Laplace domain. 3.2.2 Case B: ∀Ũs共s兲 = Us / 共s − is兲. For this case, we assume
Hence that the seabed velocity is sinusoidal and Eq. 共45兲 takes the form
dP̃s共x,s兲 cp ␣1 + ␣2 ␣3sUS
= − ssŨs共s兲 − 共 U s − U b兲 s 共43兲 W̃F共n,s兲 = − 共48兲
dx sKs s共s + i兲共s − i兲 共s − iS兲共s + i兲共s − i兲
Using zero initial term in this case, Eq. 共34兲 in this case takes the so that
冋冉 冊 冉 冊
form
␣1 + ␣2 ␣3Us e−it e it
WF共n,t兲 = 共1 − cos t兲 − i −
␣1 + ␣2 ␣3sŨs共s兲 2
2 + s − s
W̃F共n,s兲 = 共44兲
册
−
s共s2 + 2兲 共s2 + 2兲 ␣ 3 sU s i t
−i e s 共49兲
where 2 − s2
共2R − ␦兲wgh1F and
␣1 =
冋冉 冊
关共 + f 兲 − w共2R − ␦兲兴
冉
⬁
␣ˆ 1 + ␣ˆ 2 ␣ˆ 3Us e−iˆ t
兺
1
W共x,t兲 = 共1 − cos t兲 − i
cp L n=1 nˆ 2
2ˆ ˆ + s
s2␦共R − ␦/2兲 共Us − Ub兲1F
␣2 =
Ks
关共 + f 兲 − w共2R − ␦兲兴 −
eiˆ t
ˆ − s
冊 −i
␣ˆ 3sUS i t
2 e
n共ˆ − s 兲
2 册
s sin
2nx
L
共50兲
2S␦共R − ␦/2兲s1F Equations 共47兲 and 共50兲 give the corresponding transverse re-
␣3 =
关共 + f 兲 − w共2R − ␦兲兴 sponse of the pipeline with relaxed rigid porous bed model.
Hence
3.3 Analytic Solution for Rigid Porous Bed With Excita-
␣1 + ␣2 ␣3sŨs共s兲 tion of Seabed. The transverse vibration response in which the
W̃ 共n,s兲 =
F
− 共45兲
s共s + i兲共s − i兲 共s + i兲共s − i兲 seabed forcing function is the impulsive load
The solution of Eq. 共45兲 is clearly subject to the form of Ũs共s兲 F共t兲 = F0˜␦共t − t0兲
We shall assume the following forms
has the solution
冋 册
共a兲 ⬁
⌸̂ + ¯␣ˆ 2nx
兺
1
Us W共x,t兲 = 共1 − cos ˆ t兲 sin 共51兲
Ũs共s兲 = L n=1 nˆ 2 L
s
corresponding to the case Us共t兲 = H共t兲Us and whereas the response for the forcing function F共t兲 = F0 exp共i⍀̄t兲 is
共b兲 given by
兺 再 ⌸̂
nˆ 2
共1 − cos
ˆ t兲 +
¯␣ˆ
冋 ¯
ei⍀t
−
1 e−i¯t
n ˆ 2 − ⍀̄2 2ˆ ˆ + ⍀̄ 冉
Table 2 Relative density of typical geological soils
Description of soil
冊册冎
L n=1
deposit Relative density
i¯t
e 2nx
+ sin 共52兲 Very loose 0–15
ˆ − ⍀̄ L Loose 15–20
Dense 50–70
Thus by comparing the result Eq. 共51兲 with Eq. 共41兲, we can see Very dense 70–85
that the effect of an impulsive load is to increase the amplitude of 85–100
the vibration by a factor ¯␣ˆ where
¯␣ˆ = ¯␣/1F = −
F0
关共 + f 兲 − W共2R − ␦兲兴
Furthermore, when we compare the result for the case of sinu-
共53兲
⌬W共x,t兲 =
2
兺
L n=1
⬁
再冋
¯␣ˆ
¯
ei⍀t
− 冉
1 e−iˆ t
+
eiˆ t
n ˆ 2 − ⍀̄2 2ˆ ˆ + ⍀̄ ˆ − ⍀̄ 冊册冎
soidal forcing function i.e., Eq. 共52兲 with Eq. 共41兲 we see that the 2nx
effect of sinusoidal load is to increase the amplitude by the ⫻sin 共57兲
L
differential
⌬W共x,t兲 =
1
兺
L n=1
⬁
再冋 ¯␣ˆ
¯
ei⍀t
−
1 e−iˆ t
冉+
eiˆ t
n ˆ 2 − ⍀̄2 2ˆ ˆ + ⍀̄ ˆ − ⍀̄ 冊册冎 4 Analysis of Results
In this paper, we have looked at the dynamic response interac-
tion of a vibrating offshore pipeline on a moving seabed by ide-
2nx alizing the pipe as a beam vibrating on an elastic foundation so
⫻sin 共54兲 long as we have invoked the derived governing differential equa-
L tions and allowed for gradual burial of the pipe through sedimen-
tation. For the case of natural vibration, we have found that the
response is influenced by factors such as the geological character-
3.4 Analytic Solution for Deformable Porous Bed With istics of the bed, the speed of flow of oil/gas within the pipeline,
Excitation of Seabed. The transverse vibration response in which as well as the depth to the seabed. Although our analysis assumes
the seabed forcing function is inviscid fluid flow nonetheless, the dynamic interaction of the
fluid–structure–seabed is captured through the boundary condition
F共t兲 = F0˜␦共t − t0兲 Eq. 共16b兲, at the fluid–structure interface.
To enable us to interpret the significance of our results, we have
which has the solution also carried out simulations based on characteristic values of some
冋 册
of the fluid and geo–mechanical parameters that govern this fluid–
⬁
⌸̂ + 2¯␣ˆ 2nx structure–soil interaction. In particular, some of the values used in
兺
1
W共x,t兲 = 共1 − cos ˆ t兲 sin 共55兲 the simulations are listed in Tables 1–4. The values used for other
L n=1 nˆ 2 L relevant parameters for the problem treated in this report are simi-
So that comparison with Eq. 共41兲 shows that for this case the lar to those used by Olunloyo et al. 共2004兲 in a related paper 关13兴
and are as contained in Table 5. Thus, Fig. 3 illustrates the effect
factor of amplitude modulation is 2¯␣ˆ being twice what was re-
ported for the rigid porous bed, whereas the response for the forc-
ing function F共t兲 = F0 exp共i⍀̄t兲 is given by Table 3 Young’s modulus for typical geological soils
W共x,t兲 =
2
L n=1 兺
⬁
再冉 nˆ 2
冊
␣ˆ 1 + ␣ˆ 2
共1 − cos
ˆ t兲 − i 冉
␣ˆ 3US e−iˆ t
2n ˆ + s
Type of soil Young’s modulus
KN/ m2
−
eiˆ t
ˆ − s
冊 −i
␣ˆ 3sUs i t ¯␣ˆ
n共ˆ 2 − s2兲
e +
s
冋
¯
ei⍀t
Soft clay
Hard clay
Loose
1380–3450
5 865–13,800
10,350–27,600
冊册冎
n 共ˆ 2 − ⍀̄2兲 Dense sand 34,500–69,000
− 冉
1 e−iˆ t
+
eiˆ t
2ˆ ˆ + ⍀̄ ˆ − ⍀̄
sin
2nx
L
共56兲 Table 4 Typical value of drained angle of friction for sands
and silts
and comparison with Eq. 共50兲 gives the amplitude differential Sand: rounded grams
Loose 27–30
Table 1 Coefficient of permeability of typical geological soils Medium 30–35
Dense 35–38
Soil type cm/s
Sand: angular grams
Soil type 共deg兲
Clean gravel 1.0–100
Coarse sand 1.0–0.001 Loose 30–35
Fine sand 0.01–0.001 Medium 35–40
Silty 0.001–0.0001 Dense 40–45
Clay Less than Gravel with some sand 34–48
0.000001 Silts 26–35
Journal of Offshore Mechanics and Arctic Engineering MAY 2007, Vol. 129 / 113
冦
of deformation MN/m 共hard bed兲
11 Length of the F0
L 6m ¯␣ˆ = ; for rigid porous bed
pipeline 关共 + f 兲 − 共2R − ␦兲兴
12 External Do 3.5, 6.5, 12.5, 24.5 in.
diameter Y= 2F0
13 Internal Di 3, 6, 12, 24 in. 2¯␣ˆ = ; for deformable porous
diameter 关共 + f 兲 − 共2R − ␦兲兴
14 Inner radius of R Di / 2 bed
the pipeline
15 Moment of I — 共58兲
inertia
16 Uniform fluid flow velocity U 5 , 10, 15, 20 m / s
as derived from Eq. 共51兲.
through the pipe As expected, sediment cover helps to control the effect of vi-
17 Transverse pipe W W共x , t兲 bration and justifies the popular practice of burying pipes in the
displacement seabed where the geology permits 共see Figs. 5–7兲.
18 Height of ␦ Fraction
For soft seabeds, the use of increased flow velocity to dampen
sediment layer of pipe
on the pipe external diameter excitation is less effective at lower sediment cover than at the
19 Seabed Ks 0.02 higher level as can be seen from Figs. 6 and 7 compared to Figs.
permibility 4 and 5. Such increased pumping speed, however, transfers the
coefficient likelihood of rupture from one of shear to that of fatigue since, as
20 Wave number K 0.1 seen from the Fig. 8–11, as it is usually accompanied by increased
21 Porosity cp 0.03
coefficient or
frequency in the case of the hard seabed.
index However, the optimal damping is not linearly dependent on
internal fluid velocity for the soft bed neither is it inversely de-
pendent for the hard ocean floor.
This means that the optimal policy for the effective manage-
ment of the pipeline system will require adequate monitoring of stays within reasonable limits. One of such domains occurs at low
the sediment cover. In subsequent publications, this and other is- to moderate pumping speeds, while the other requires relatively
sues will be discussed further 共see Figs. 12–18兲. high and unrealistic velocities for the physical model under ex-
Being at the bottom of the sea, drag effects as usually calcu- amination as our solution must also satisfy the requirement that
lated by use of Morison empirical formula are of second order the velocity of the conveyed fluid remains within reasonable
consideration especially in cases such as ours where the pipe is limits.
laid along the flow direction. However, such contribution can be By examining any of Figs. 19–24, it can be deduced that the
significant where the inclination between the fluid flow and pipe range of the optimal domain for the fluid conveyance velocity
direction is appreciable and of course, when dealing with risers. increases, in general, with the level of sediment cover. Further-
Where the pipe is laid at an inclination, to the fluid flow direction, more, comparison of the plots for soft and hard seabeds 共e.g., by
the appropriate governing differential equation to be solved was way of comparison of Figs. 19 and 20, 21 and 22 etc.兲 shows that
given by us in Ref. 关13兴 while the more general problem of the the optimal domain is on the average wider for a hard bed than for
effect of a nonstationary seabed on the Morison hydrodynamic a soft bed and that even this range can be enlarged by increasing
force was addressed by Olunloyo and Osheku in Ref. 关14兴. the size of the conduit pipe as is evidenced from Fig. 24 when
However, in this paper, we have also attempted to examine the compared with Figs. 22 or 20 for example.
level of attenuation of the deflection in relation to the sediment On the other hand we also find that with increasing pipe radius,
cover ␦ and the pumping speed. In this regard, we find that in the amplitude of vibration increases. This can be confirmed by
general, the amplitude of vibration dies progressively with in- examining sequentially Figs. 13, 5, 15, and 17 corresponding to
crease in sediment cover ␦. With respect to pumping speed, the the deflection pattern for pipes of radius 75 mm, 150 mm, 300
relation is nonlinear and may in fact be parabolic. mm, and 600 mm, respectively, on a soft seabed while Figs. 14,
For example, in Figs. 19–24 we find that for each of the profiles 10, 16, and 18 display the corresponding patterns for pipes laid on
there are at least two regions where the maximum pipe deflection a hard seabed. However, irrespective of the seabed geology or
Fig. 6 Dynamic interaction response profile for soft seabed Fig. 8 Dynamic interaction response profile for hard seabed
and pipe internal diameter 150 mm „6 in. … at 80 s and pipe internal diameter 150 mm „6 in. … at 80 s
Journal of Offshore Mechanics and Arctic Engineering MAY 2007, Vol. 129 / 115
internal fluid velocity, as the pipe size increases the dynamic re-
sponse converges to the bimodal profile as can be confirmed from
Figs. 17 and 18.
As expected, external excitation will increase the response of
these pipes for which an amplification factor has been derived for
the case of impulsive loading as earlier highlighted in Eq. 共58兲.
These results are of contemporary interest in the oil/gas industry
where deep sea exploration is now receiving significant attention.
Fig. 11 Dynamic interaction response profile for hard seabed Fig. 13 Dynamic interaction response profile for soft seabed
and pipe internal diameter 150 mm „6 in. … at 80 s and pipe internal diameter 75 mm „3 in. … at 80 s
Fig. 15 Dynamic interaction response profile for soft seabed Fig. 18 Dynamic interaction response profile for hard seabed
and pipe internal diameter 300 mm „12 in. … at 80 s and pipe internal diameter 600 mm „24 in. … at 80 s
Journal of Offshore Mechanics and Arctic Engineering MAY 2007, Vol. 129 / 117
Fig. 22 Maximum transverse response as a function of flow Fig. 24 Maximum transverse response as a function of flow
velocity for hard seabed and pipe internal diameter of velocity for hard seabed and pipe internal diameter of
300 mm „12 in. … 600 mm „24 in. …
Journal of Offshore Mechanics and Arctic Engineering MAY 2007, Vol. 129 / 119