Quasi Static Moooring Analysis

You might also like

Download as pdf or txt
Download as pdf or txt
You are on page 1of 23

Ocean Engineering 29 (2002) 627–649

http://www.elsevier.com/locate/oceaneng

Semi-analytical quasi-static formulation for


three-dimensional partially grounded mooring
system problems
Y.T. Chai, K.S. Varyani *, N.D.P. Barltrop
Department of Naval Architecture and Ocean Engineering, University of Glasgow, Glasgow G12 8QQ,
Scotland, UK

Received 19 October 2000; accepted 10 January 2001

Abstract

A semi-analytical quasi-static formulation based on the catenary approach capable of solving


three-dimensional partially grounded and fully suspended multi-leg mooring-system problems
is presented in this paper. The advantage of the present formulation is that only a compact
governing vector equation, derived in terms of grounded cable length and anchor tension vec-
tor, need to be solved for each mooring cable to determine its grounded and suspended body
response, and only a small number of discrete segments are needed to predict its behaviour
accurately due to its inherent slope continuity. In addition, it is capable of handling arbitrarily
inclined seabed interaction effects, varying cross-sectional and material properties as well as
external attachment objects. The generality of the present formulation allows quick parametric
analysis of different forms of multi-leg mooring system configurations as well as different
types of flexible riser systems to be carried out.  2002 Elsevier Science Ltd. All rights
reserved.

Keywords: Semi-analytical; Partially grounded; Mooring system

1. Introduction

Mooring systems are widely employed by floating offshore structures for station
keeping purposes and can be classified either as single-point mooring systems or as
spread mooring systems. Single-point mooring is used mainly for ship-shaped vessels

* Corresponding author.
E-mail address: k.s.varyani@eng.gla.ac.uk (K.S. Varyani).

0029-8018/02/$ - see front matter  2002 Elsevier Science Ltd. All rights reserved.
PII: S 0 0 2 9 - 8 0 1 8 ( 0 1 ) 0 0 0 3 8 - 5
628 Y.T. Chai et al. / Ocean Engineering 29 (2002) 627–649

Nomenclature
s, se Unstressed and stretched cable arc length
A, Aatt Cross-sectional area of cable and attachment object
AT, AL Transverse and frontal wind area of floater above mean waterline
Ls, Ts Length and draft of floater
fc Form factor of factor
e Axial strain
A Cross-sectional area of cable
E Young’s modulus
D Unstressed cable diameter
L Unstressed cable length
rsw, rair Water and air density
Cdn, Cdt Normal and tangential drag coefficient of cable segment
n Kinematic viscosity of seawater
Cd Drag coefficient of attachment object
Te Effective tension
w Wave frequency
U0 Unstressed grounded cable length
t Scaling factor for anchor tension and grounded length correction
tb Scaling factor for floater displacement correction
Swet Wetted surface area of floater
V̄c, V̄wd Mean current and wind speed
q⬘, j⬘, y⬘ Seabed slope Eulerian angles
Cdcy, Cdcy Sway and yaw current drag coefficients
Cdwx, Cdwy, Cdw Surge, sway, and yaw wind drag coefficients
Qx, Qy, Qy Surge, sway and yaw quadratic transfer functions
qc, qwd, qwv Current, wind and wave angle of attack
N, Ng Number of cable and grounded segments respectively
NL Number of mooring legs
→ →
R, R L Position vector with respect to global and seabed reference system

uc Current velocity vector
→ → →
w g, w m, w c Seabed reaction, current loading and wet weight of cable per
unit length

F att
<
Concentrated force due to external attachment objects
J Tension vector at seabed anchor point
[⌫lg] Transformation matrix between the global reference system and the
local seabed reference system
[⌫bg] Transformation matrix between the body-fixed and global reference
system
Y.T. Chai et al. / Ocean Engineering 29 (2002) 627–649 629

[K] Cable restoring stiffness matrix


Tm具b典 Top-end loading vector due to the bth mooring cable

P o具b典 Initial position vector of the bth fairlead connection point with
respect to the local seabed reference system

r 具b典 Position vector of the bth fairlead connection point with respect to
the body-fixed reference system

F Steady environmental loading vector

x Displacement vector of floater

while spread mooring is used mainly for semi-submersibles where environmental


loading acting on the platform is relatively insensitive to direction. These systems
usually consist of an array of multi-component cables that are partially grounded on
the seabed with their bottom-end point fixed to the seabed by means of anchors and
their top-end point connected to the floater either through a turret system or through
the windlass. The special characteristics of these mooring cables are that they are
totally flexible (ideal case) and are incapable of resisting compression loading. These
features thus result in the highly non-linear geometry of the mooring system, which
is not known a priori. Special methods are therefore required for their analysis.
Although dynamic effects are important and are within the capability of existing
software, quite often only quasi-static analysis is needed which is the main emphasis
of this paper, especially in the preliminary design stage where quick parametric
analysis is required without having to resort to the more time-consuming dynamic
analysis.
The frequently used methods for the three-dimensional quasi-static analysis of
mooring cable systems include the lumped-parameter (mass) method (Leonard and
Nath, 1981), the standard finite-element technique (Webster, 1975), the cable
element method (Peyrot and Goulios, 1979; Peyrot, 1980), the direct integration
technique (Leonard, 1979; Chiou and Leonard, 1991) and the method of imaginary
reaction (Skop and O’Hara, 1970, 1972; Skop, 1988). The first three methods are
finite-element based methods using straight-link or cable elements to model the
cable system. Substantial memory storage and computational time can be con-
sumed in solving the large set of simultaneous equilibrium equations to obtain the
unknown nodal displacements of these elements. The fourth method, however,
relies on the numerical integration of the governing differential equations along
each cable length to solve the system as an iterative set of initial-value problems,
while the fifth method employs the flexibility technique to solve the system as a
non-linear boundary value problem. In the latter method, unknown reactions are
first prescribed at the cable terminal ends and then iterated upon numerically until
the specified end boundary conditions are satisfied. Although these methods are
relatively more efficient and do not require extensive use of memory space, sig-
630 Y.T. Chai et al. / Ocean Engineering 29 (2002) 627–649

nificant numbers of integration points will still be needed in order to predict the
cable behaviour accurately. A more desirable approach is therefore to use the
semi-analytical technique, as adopted by Huang and Vassalos (1993). However,
this formulation only applies to the fully suspended single cable problem, render-
ing it inadequate for the analysis of the partially grounded mooring system prob-
lem. In addition, the formulation does not permit slope continuity to be maintained
throughout the cable length.
An alternative semi-analytical formulation based on the catenary model
(O’Brien, 1967) is presented in this paper. The advantage of using catenary-based
formulation lies in its inherent incompressibility characteristic and its slope conti-
nuity feature that allows a smaller number of discrete segments to be used to
model the cable structure accurately, as compared to other models using straight-
link elements. The three-dimensional cable element method mentioned earlier is
actually based on such a model to take advantage of these properties. The merit
of the present three-dimensional formulation over other formulations is that it
allows solution of both the partially grounded and fully suspended mooring system
problems in a combined manner using only a vector (three) equation to be solved
for each mooring cable. In addition, it is capable of handling arbitrarily inclined
flat seabed interaction effects, varying cross-sectional and material properties as
well as external attachment objects (spring buoys or clump weights). Its generality
coupled with its compactness therefore allows efficient analysis to be carried out
for different forms of multi-leg mooring system configurations. Although bending
stiffness effect is ignored, its application for the preliminary analysis of flexible
riser systems (such as lazy-wave, free hanging, steep-s and steep-wave riser
configurations) is still valid, as it has no significant effect on the global response
except only at regions of high curvature. The formulation of this technique and
its solution approach for a single cable partially grounded on the inclined seabed
is first given in Section 2. Extension of the technique to solve multi-leg mooring
system problems is then described in Section 3. Numerical examples demonstrat-
ing the capability of the formulation are finally presented in Section 4.

2. Mathematical formulation

The basic assumptions in the present modelling of mooring cables are as follows:

1. Flat frictionless seabed


2. Fixed anchor point
3. Non-negative tension
4. Zero bending and torsional stiffness
5. Non-submergence of the grounded cable segment into seabed surface

The cable is first discretized into a number of major segments such that the cable
properties (Young’s modulus, area, weight, drag coefficients) in each of these seg-
ments are constant. These major segments are then further sub-divided into a number
Y.T. Chai et al. / Ocean Engineering 29 (2002) 627–649 631

Fig. 1. Partially grounded mooring cable configuration.

of smaller segments such that the position- and orientation-dependent drag loading
can be accurately applied as uniform loading acting on each smaller segment. As
shown in Fig. 1, two reference systems are employed. They are the O-XYZ global
reference system and the O-xyz local seabed reference system. The local seabed
reference system is obtained by rotating the global O-XY plane until coincidence
with the inclined seabed plane is achieved. The relation between these two reference
systems is expressed as follows:
→ →
R L⫽[⌫lg]R
where

[⌫lg] ⫽

冤 冥
cos(q⬘)cos(y⬘) cos(q⬘)sin(y⬘) −sin(q⬘)
sin(q⬘)cos(y⬘)sin(j⬘)−cos(j⬘)sin(y⬘) sin(q⬘)sin(y⬘)sin(j⬘)+cos(j⬘)sin(y⬘) cos(q⬘)sin(j⬘)
sin(q⬘)cos(y⬘)cos(j⬘)+sin(j⬘)sin(y⬘) sin(q⬘)sin(y⬘)cos(j⬘)−sin(j⬘)cos(y⬘) cos(q⬘)cos(j⬘)

and q⬘, j⬘, y⬘ are the Eulerian angles defining the slope of the inclined seabed plane.
632 Y.T. Chai et al. / Ocean Engineering 29 (2002) 627–649

2.1. Force balance relation

The force equilibrium relation for the kth segment with respect to the local seabed
reference-system is then expressed as follows:



∂R →
k⫺1
→ → → → →
Te(s)k ⫽J ⫺ w i(si+1⫺si)⫺F attk−1⫺g f(s)⫺(w k⫺w gkm(s))(s⫺l(s)k) (1)
∂s i⫽0,i⫽N g


where w k is the net uniform loading acting on a discrete segment:
→ → → →
w k⫽w mkⴙw gkⴙw ck


w mk is the mean external fluid loading due to current and is calculated using the
drag term of Morrison’s equation. This drag loading is dependent on the cable
geometry itself, which is not known a priori and thus has to be determined in an
iterative manner using the tangential vector of its chord segment as the representative
slope of the cable segment.

→ 1
w mkⴝ rswDk 1⫹e
2 冉 冉 冊 冊冉 | 冉 冊 | 冉 冉 冊 冊
sk+sk+1
2 k
Cdtk →
uc
sk+sk+1 → → sk+sk+1 → →
2
· t k uc
2
·tk t k (2)

⫹Cdnk
| 冉 冉 冊冊 | 冉 冉 冊冊 冊
→ →
t k⫻u c
sk+sk+1
2

⫻t k


t k⫻u c
sk+sk+1
2

⫻t k


in which t k is the tentative tangential vector of its chord segment defined by:
→ →
→ R(sk+1)j −R(sk)j
t k⫽ → →
||R(sk+1)j −R(sk)j ||


w gk is the seabed reaction loading expressed as follows:


w gk⫽ 再 兵0
兵0
0 −wcz −wmz 其T kⱕNg
0 0其T
k k

otherwise
(3)


F attk is the total concentrated loading up to the kth segment due to external objects
(spring buoy or clump weight) being attached along its length expressed in the fol-
lowing form independent of cable orientation.
→ 1
F attk⫽ rsw
2 冉冘
i⫽0
k

CdiAatti||→
u c(si+1)||→
u c(si+1)⫹W atti

冊 (4)
Y.T. Chai et al. / Ocean Engineering 29 (2002) 627–649 633

<

gf(s) is the external force vector acting on the last grounded cable segment
given by:

冦
wxNg(sNg+1−sNg)


wyNg(sNg+1−sNg)
(wmz +wczNg)(sNg+1−U0)
Ng
冧 s⬎sNg+1


wxNg(U0−sNg)

g f(s)⫽  wyNg(U0−sNg) sNg+1ⱖs⬎U0

冦 冧
(5)
0

0
0 otherwise
0

The auxiliary functions l(s)k and m(s) are defined as:

l(s)k⫽ 再
U0 sNg+1ⱖs⬎U0
sk otherwise
(6)

m(s)⫽ 再1 sNg+1ⱖs⬎U0
0 otherwise
(7)

2.2. Compatibility and material constitutive relations

The two additional relations that have to be satisfied by the cable are:

앫 Compatibility relation:

→ →
∂R 1 ∂R
| ||
∂se

1+e ∂se
⫽1
| (8)
634 Y.T. Chai et al. / Ocean Engineering 29 (2002) 627–649

From this relation the magnitude of the effective tension is derived direct from Eq.
(1) as:

| 冘
k⫺1
→ → →
Te(s)k⫽ J ⫺
i⫽0,i⫽Ng
→ → →
w i(si+1⫺si)⫺F attk−1⫺g f(s)⫺(w k⫺w gkm(s))(sⴚl(s)k)
| (9)

앫 Linear material constitutive relation:

Te(s)k⫽EkAke(s)k (10)

For non-linear cable material such as synthetic fibre rope, the elastic modulus
becomes stress dependent and this will complicate the subsequent formulation.
To simplify the analysis the tangent modulus is assumed constant throughout the
sub-segment length determined based on the tentative tension acting at its mid-
point as shown below. This approach avoids the application of the non-linear
material constitutive relation and allows the simple linear relation to be used
instead.

冉 冊
冢 冣
sk+sk+1 1−v
Te
1 2
Ek⫽
k

vC Ak

where C and v are material constants determined experimentally.

2.3. Combined grounded and suspended cable geometry

The expression of the combined grounded and suspended cable geometry is then
readily obtained by utilising relations (8) and (9) in Eq. (1) and carrying out analyti-
cal integration of the resulting vector equation with respect to the unstressed arc
length variable, s.


k⫺1
→ → → →
R(s)⫽ ⌬R(si+1)i⫹⌬R(U0)Ngm(s)⫹⌬R(s)k (11)
i⫽0

where:
Y.T. Chai et al. / Ocean Engineering 29 (2002) 627–649 635


⌬R(s)k⫽ (12)

冉 冊
冉
→ s−l(s)k Q(s)k−Q(l(s))k →
g (l(s)k) + − ||w k||⫽0
EkAk → →
||w k−w g Ngm(s)||


→ → →
 s −l(s)k+(w k−w g Ngm(s))· g (s)k(Q(s)k−Q(l(s)k)k)+Te(s)k−Te(l(s)k)k (w
2 2
→ →
k−w g Ngm(s))


2EkAk → 3 → 2 → →
||w k−w g Ngm(s)|| ||w k−w g Ngm(s)||


(s−sk) 冉 1
+
1 →
Te(s)k EkAk
g (sk) 冊 →
||w k||=0

冢 冣
→ → 2 → → →
s||w k−w gNgm(s)|| −(w k−w g Ngm(s))· g (s)k
Q(s)k⫽sinh ⫺1
(13)
冑|| g (s) || ||w −w

k
2 →
k

g Ng
2 → →
m(s)|| −((w k−w g Ngm(s))· g (s)k)

2


N⫺1
→ → → → → →
g (s)k⫽J⫺ w i(si+1⫺si)⫺F attk−1⫺g f(s)⫹(w k⫺w gkm(s))l(s)k
<

(14)
i⫽0,i⫽Ng

2.4. Solution technique for a single cable problem

Given the known position of its two end-points the non-linear boundary value
problem is completely defined and is solved using iterative incremental correction
technique.


N⫺1
→ → → →
P 0⫽ ⌬R(si+1)i⫹⌬R(U0)Ng⫹R(0) (15)
i⫽0

→ →
in which R(0) is the position vector of its fixed anchor point and P 0 is the prescribed
position vector of its top-end point.
Incremental correction:

∂fx ∂fx 1 ∂fx

冤 冥
−1

∂Jx ∂Jy s ∂⍀

冦 冧 冦冧
⌬Jx fx
∂fy ∂fy 1 ∂fy →
⌬J⫽ ⌬Jy ⫽⫺t fy ⫽⫺t[K] f
<

(16)
∂Jx ∂Jy s ∂⍀
s⌬⍀ fz
∂fz ∂fz 1 ∂fz
∂Jx ∂Jy s ∂⍀

where f is the miss-closure vector function defined as:
636 Y.T. Chai et al. / Ocean Engineering 29 (2002) 627–649


N⫺1
→ → → → → →
f (U0,J )⫽ ⌬R(si+1)i⫹⌬R(U0)Ng⫹R(0)⫺P 0 (17)
i⫽0

in which t is the scaling factor introduced to allow convergence of the solution. The
differentiating variable ⍀ and the value of the parameter s in the above matrix
depend on the state (partially grounded or fully suspended) of the cable.

앫 In fully suspended state Jzⱖ0, U0=0: ⍀⬅Jz, s=1


앫 In partially grounded state U0⬎0, Jz=0: ⍀⬅U0, s=wzNg⫺wgz
Ng

The onset of grounding occurs when Jz⬍0 while the transition from partial grounding
to full suspension occurs when U0⬍0. The restoring stiffness matrix [K] is then
determined analytically assuming that the external loading is constant. Alternatively,
the matrix can be evaluated numerically using the approximate formula derived by
Broyden (1965). This formula does not require expensive function evaluations except
only in the initial approximation stage and works particularly well in most cases
when drag loading is present.
→ → → →
(⌬J j−1−[K]j−1( f j − f j−1))丢⌬J j−1
[K]j ⫽[K]j−1⫹ → → →
[K]j−1 (18)
J j−1·([K]j−1( f j − f j−1))

Iteration is terminated when specified convergence tolerance is achieved. With the


known anchor tension and grounded length, the response of the entire cable in the
local seabed reference system is completely defined. The cable geometry in the global
reference system is then readily obtained by simple matrix transformation.

3. Multi-leg mooring system analysis

The solution technique described in the previous section can now be applied to
determine the quasi-static response of multi-leg mooring cable systems using an
iterative procedure described in the following section. The system is assumed to
consist of NL mooring legs connected to the floater through the turret system as
shown in Fig. 2. Only the dominant modes of motion (i.e. surge, sway, and yaw)
are considered. Although in practice the turret structure is free to rotate about its
own axis, its rotation has no significant effect on the quasi-static response due to
the small chaintable radius and is therefore neglected.
The equilibrium condition of the moored floater with respect to the body-fixed
co-ordinate system, 0⫺xsys is now expressed as follows:


NL
→ →
T m具b典⫽⫺F (19)
b⫽1
Y.T. Chai et al. / Ocean Engineering 29 (2002) 627–649 637

Fig. 2. Plan view of turret-moored FPSO.


Where F is the total steady environmental loading acting on the floater and Tm具b典 is
the restoring load due to the bth mooring leg determined from the following con-
straint relation:

冦冧
x1

N⫺1
→ → → →
R(s ) ⫹⌬R(U ) ⫹R(0) ⫽P ⫹[⌫lg] x2 ⫺([⌫lg][⌫bg]
具b典
i+1 i
具b典
0 Ng
具b典 具b典
0 (20)
i⫽0
0


⫺[⌫lg]) r 具b典 b⫽1,2…NL
where [⌫bg] is the transformation matrix between the global and the body-fixed refer-
ence system defined by:

冤 冥
cos(x3) −sin(x3) 0
[⌫bg]⫽ sin(x3) cos(x3) 0
0 0 1
638 Y.T. Chai et al. / Ocean Engineering 29 (2002) 627–649

Eqs. (19) and (20) thus form a set of vector equations describing the complete steady-
state behaviour of the mooring-floater system.

3.1. Environmental loading on floater

The total steady environmental loading acting on the floater consists of three
components namely current, wind and wave drift loading components. The steady
current and wind loading are usually estimated empirically using the drag-coefficients
published by OCIMF (1994) or derived directly from experiments, while the steady
wave drift loading is normally determined theoretically using three-dimensional
potential theory. The expressions of these steady environmental loading components
are given as follows:
Current loading:

0.075||cos(qc−x3)||cos(qc−x3)

冦 冉 冉 冊冊

S (1+fc)
Ls||V̄ccos(qc−x3)|| 2 wet
→ 1 log −2
F c(x3)⫽ rswV̄2c n (21)
2
Cdcy(qc−x3)LsTs
Cdcy(qc−x3)L2sTs

Wind loading:

冦 冧
Cdwx(qwd−x3)AT
→ 1
F wd(x3)⫽ rairV̄2wd Cdwy(qwd−x3)AL (22)
2
Cdwy(qwd−x3)ALLs

Wave loading:

冦 冧
Qx(w,qwv−x3)



F wv(x3)⫽2 S␵␵(w) Qy(w,qwvx3) dw (23)
0 Qy(w,qwv−x3)
Y.T. Chai et al. / Ocean Engineering 29 (2002) 627–649 639

3.2. Solution technique

The initial position of the floater is first estimated and the equilibrium configur-
ation of each mooring leg is then determined using the procedure described in the
previous section. The floater equilibrium relation will not be satisfied in the first
iteration cycle and a new estimate of the floater displacement will have to be made.
This is achieved by updating the tentative floater displacement with the following
displacement correction:

冉冘 冊 冉冘 冊
NL → → NL
→ ∂T m具b典 ∂F −1 → →
⌬x j ⫽⫺tb →
⫹ → T m具b典⫹F (24)
b⫽1 ∂x ∂x j b⫽1 j

With the known updated floater displacement, the miss-closure between the top-
end point of each mooring leg and its fairlead connection point is now determined
as follows:

冦冧
x1

N⫺1
→ → → → → →
f (U0,J )⫽ R(s ) ⫹⌬R(U ) ⫹R(0) ⫺P ⫺[⌫lg] x2
具b典
i+1 i
具b典
0 Ng
具b典 具b典
0 (25)
i⫽0
0
j


⫺([⌫lg][⌫bg]j ⫺[⌫lg]) r 具b典 b⫽1,2…NL

Improved estimate of the grounded cable length and anchor tension vector for
each mooring leg is then readily obtained from Eq. (16) and their revised top-end
tensions are again applied to the floater. The procedure is repeated until the miss-
closure distance of each leg and the load imbalance of the floater reaches a specified
convergence tolerance. The advantage of this approach over the finite-element tech-
nique is that the large memory requirement needed for the solution of a large set of
simultaneous equations is avoided. Effectively only a compact vector equation needs
to be solved for each mooring leg.

4. Numerical examples

An object-orientated mooring-system analysis program (SEMIF98) based on the


above formulation has been developed in C++. This program is capable of analysing
the quasi-static response of coupled mooring-floater systems under different environ-
mental loading combinations. A symmetric mooring system comprising 10 multi-
component cables connected to a FPSO is then analysed using the program. The
640 Y.T. Chai et al. / Ocean Engineering 29 (2002) 627–649

Table 1
Particulars of FPSO

Designation Symbol

Length Ls 240.0 m
Breadth Bs 44.0 m
Depth Ds 32.0 m
Draft Ts 20.0 m
Underwater hull area Swet 18,315 m2
Wind side area AL 5760 m2
Wind frontal area AT 1560 m2
Form factor fc 0.4
Center of Turret from FP 24.0 m

objective is to demonstrate the capability of the present formulation only rather than
to apply it for detailed analysis or design. For simplicity sake, the FPSO is modelled
as an elliptical hull cylinder moored in a water depth of 404 m (Table 1) subjected
to oblique current and wind loading only (Table 2). Each mooring leg is assumed
to be made up of 4 segments (wire–chain–wire–wire) with one spring buoy attached
as given in Table 3. Both horizontal and inclined seabed conditions are considered
in the examples including a combined mooring-riser system example.

4.1. Horizontal seabed example

The steady-state configuration of the mooring system partially grounded on the


horizontal frictionless seabed subjected to oblique environmental loading is shown
in Figs. 3 and 5. The grounded length, the anchor and fairlead tension for each
mooring leg is given in Table 4 while the tension distribution along each leg is
shown in Fig. 4. To validate the accuracy of the present formulation, these results
are compared with those obtained using the lumped-parameter (mass) method, which
has also been developed by the authors. Good agreement between the two methods

Table 2
Environmental data

Designation Symbol

Current: V=V̄b+(V̄c⫺V̄b)(z/Dw)n
Surface speed V̄c 1.20 m/s
Seabed speed V̄b 0.40 m/s
Index n 1
Direction qc 30°
Water depth Dw 404 m
Wind
Mean speed V̄wd 43 m/s
Direction qwd 10°
Y.T. Chai et al. / Ocean Engineering 29 (2002) 627–649 641

Table 3
Particulars of the mooring cable

Designation:
Length of mooring cable 2100 m
Chain table above seabed 384 m
Pre-tension 527.7 kN
Pre-tension angle 36.35°
Segment 1: Steel wire
Length form turret 200 m
Diameter 5.5 in.
Mass 83 kg/m
Axial stiffness 1150 MN
Breaking strength 15.99 MN
Spherical buoy properties:
Diameter 4.68 m
Total dry weight 206 kN
Net buoyancy 334 kN
Segment 2: Steel wire
Length 400 m
Diameter 5.5 in.
Segment 3: Chain
Length 400 m
Diameter 5 in.
Mass 353 kg/m
Axial stiffness 1390 MN
Breaking strength 11.516 MN
Segment 4: Steel wire
Length 1100 m
Diameter 5.5 in.

Fig. 3. Profile view of the mooring system in oblique environment loading.

is noted in Table 5. The lumped-parameter model requires about 71 straight link


elements while only 20 segments are needed by the present formulation to achieve
an accuracy level within 0.25%.

4.2. Inclined seabed example

The seabed is assumed inclined in this numerical example. The Eulerian angles
defining the slope of the seabed plane are taken as q⬘=⫺4°, y⬘=25° and j⬘=0°. The
642 Y.T. Chai et al. / Ocean Engineering 29 (2002) 627–649

Fig. 4. Effective tension distribution along each mooring leg in oblique environmental loading.

Table 4
Horizontal seabed condition

Leg Uo (m) Top-end tension (N) Bottom-end tension (N)

1 1325.04 995,744 1,179,414


2 1325.48 993,029 1,174,918
3 1370.78 726,838 872,300
4 1421.50 445,535 549,704
5 1449.98 290,638 374,003
6 1462.11 223,674 301,530
7 1462.49 220,950 300,203
8 1451.04 280,811 367,333
9 1420.31 444,834 554,609
10 1369.04 732,241 883,171

Fig. 5. Plan view of the mooring system subjected to oblique environmental loading.
Y.T. Chai et al. / Ocean Engineering 29 (2002) 627–649 643

Table 5
Comparison between the lumped-parameter model and the present formulation

Legs Top-end tension (N)

Lumped-parameter model Present formulation


(71 segments) (20 segments)

1 1,177,658 1,179,414
2 1,172,671 1,174,918
3 871,243 872,300
4 549,155 549,704
5 374,488 374,003
6 302,261 301,530
7 301,061 300,203
8 367,826 367,333
9 554,135 554,609
10 883,512 883,171
Equilibrium position
Surge 20.10 m 20.08 m
Sway 17.11 m 17.10 m
Yaw 6.25° 6.25°

anchor point of each mooring leg is assumed to lie on the same vertical axis before
and after the seabed rotation. To illustrate the effect of the inclined seabed plane on
the mooring configuration, environmental loading is neglected in the analysis. As
expected (Figs. 6–8), the grounded cable segments respond to the seabed inclination

Fig. 6. Plan view of the mooring system subjected to an inclined seabed.


644 Y.T. Chai et al. / Ocean Engineering 29 (2002) 627–649

Fig. 7. Profile view of the mooring system subjected to an inclined seabed.

Fig. 8. Side view of the mooring system subjected to an inclined seabed.

by sliding downwards on the inclined seabed plane resulting in an increase in the


anchor tension of the upward lines (Table 6). In practice, the seabed friction will
reduce the sliding down effect of the mooring lines by forming some form of groove
on the seabed. However, in the authors’ opinion, the persistent top-end excitation of

Table 6
Inclined seabed condition

Floater equilibrium position

Surge 26.150 m
Sway 9.516 m
Yaw 0°

Legs Uo (m) Anchor tension (N) Top-end tension (N)

1 1397.43 704,856 694,905


2 1397.82 706,007 692,475
3 1393.16 679,320 721,704
4 1390.09 609,279 741,063
5 1393.75 501,632 717,933
6 1406.39 378,600 638,609
7 1407.86 367,740 629,476
8 1394.75 487,966 711,635
9 1390.07 599,010 741,187
10 1392.63 673,765 725,073
Y.T. Chai et al. / Ocean Engineering 29 (2002) 627–649 645

Table 7
Particulars of steel riser

Designation:
Horiz. distance 1450 m
Vert. distance 384 m
Upper segment:
Length 736 m
Outer diameter 0.254 m
Wall thickness 0.0254 m
Axial stiffness 3830.7 MN
Buoyancy segment:
Length 390 m
Outer diameter 0. 675 m
Riser diameter 0.254 m
Wall thickness 0.0254 m
Axial stiffness 3830.7 MN
Lower segment:
Length 516 m
Outer diameter 0.254 m
Wall thickness 0.0254 m
Axial stiffness 3830.7 MN
Internal fluid 840 kg/m3

the mooring system will likely cause the mooring lines to fully slide downwards on
the inclined seabed plane eventually. This may therefore justify the negligence of
the seabed frictional effect on the steady-state response of the mooring system for
the initial design stage.

4.3. Combined mooring-riser system example

For this example, the steady-state response of a combined mooring-riser system


is carried-out. The mooring layout adopted is a 3×3 bundle system configuration
with three low long-wave steel risers being hanged off between the line bundles.
The separation angle of each mooring line is taken as 3°. Spring buoy is removed
from each mooring line to prevent collision due to the small separation angle. The
particulars of the steel risers are given in Table 7. Although the bending stiffness
effect is ignored, there will not be a significant change to the steady response of the
FPSO due to the large length of the steel riser, which effectively causes it to behave
like a cable globally. Negligence of the bending stiffness effect will therefore be
rational and justifiable for the initial design stage. The steady-state deformed con-
figuration of the combined mooring-riser system subjected to the same oblique
environmental loading is now illustrated in Figs. 9 and 10 while the equilibrium
position is given in Table 8.
646 Y.T. Chai et al. / Ocean Engineering 29 (2002) 627–649

Fig. 9. Profile view of the mooring-riser system in oblique environmental loading.

Fig. 10. Plan and side view of the mooring-riser system configuration in oblique environmental loading.

Table 8
Steady drift of combined mooring-riser system

Equilibrium position:

Surge 17.01 m
Sway 19.60 m
Yaw 6.25°

5. Conclusions

A semi-analytical quasi-static formulation based on the catenary approach capable


of solving three-dimensional partially grounded and fully suspended multi-compo-
nent mooring-system problems has been presented. The advantage of the present
formulation lies in its compactness that requires only a vector (three) equation to be
solved for each mooring leg to determine both its grounded and suspended body
Y.T. Chai et al. / Ocean Engineering 29 (2002) 627–649 647

response, and only a small number of discrete segments are needed to model the
cable behaviour accurately due to its inherent slope continuity feature. In addition,
it is capable of handling arbitrarily inclined flat seabed interactions as well as external
attachment objects. Different forms of multi-leg mooring systems as well as different
types of flexible riser system configurations can therefore be analysed without requir-
ing extensive use of memory storage and computational effort as compared to the
finite-element technique.

Appendix A

The analytical expression of the coefficients in the restoring stiffness matrix are
given as:

冘 冉冉 冊
→ →
∂f sk+1−sk Q(sk+1)k−Q(sk)k g (sk)k ∂
N


⫽ ⫹ [I]⫹ →
(Q(sk+1)k⫺Q(sk)k)
EkAk → →
∂J k⫽1,k⫽Ng ||w k|| ||w k|| ∂J


→ →
→ w k· g (sk)k ∂ 1 ∂
⫺w k →
(Q(sk+1)k⫺Q(sk)k)⫹ →
(T(sk+1)k⫺T(sk)k) (A1)
→ → 2
||w k|| ∂J ||w k|| ∂J


Q(sk+1)k−Q(sk)k→ T

||w k||
wk ⫹
ENgANg


冊冊 冉
sNg+1−sNg Q(sNg+1)Ng−Q(U0)Ng

||w Ng−w g Ng||



Q(U0)Ng−Q(sNg)Ng g (sNg)Ng ∂
⫹ →
[I]⫹ →
(Q(sNg)Ng⫺Q(U0)Ng)

||w Ng|| ||w Ng|| ∂J


→ → →
g (U0)Ng ∂ → w Ng· g (sNg)Ng ∂
⫹ →
(Q(sNg+1)Ng⫺Q(U0)Ng)⫺w Ng →
(Q(U0)k
→ → →
||w Ng−w g Ng|| ∂J ||w Ng|| ∂J

⫺Q(sNg)Ng)⫹

1
|| Ng||
w
2
∂J

(T(U0)Ng⫺T(sNg)Ng)⫹

Q(U0)Ng−Q(sNg)Ng→ T

|| Ng||
w
w Ng 冊
→ → →
→ → (w Ng)−w g Ng)· g (U0)Ng ∂
⫺(w Ng⫺w g Ng)( →
(Q(sNg+1)k⫺Q(U0)Ng)
→ →
||w Ng−w g Ng|| ∂J
1 ∂ Q(sNg+1)Ng−Q(U0)Ng → →
⫹ (T(sNg+1)Ng⫺T(U0)Ng)⫹

(w Ng⫺w g Ng)T)
→ → 2 → →
||w Ng−w g Ng|| ∂J ||w Ng−w g Ng||
648 Y.T. Chai et al. / Ocean Engineering 29 (2002) 627–649

冘冉 冉 冊
→ → → →
∂f sk+1−sk g (sk)k → w k· g (sk)k ∂
N

⫽ ⫺ →(sk)k⫹ ⫺w k (Q(sk+1)k⫺Q(sk)k)
∂U0 k⫽Ng⫹1 EkAk → → 3 ∂U0
||w k|| ||w k||

冉 冊冊
→ → → →
wk ∂ Q(sk+1)k−Q(sk)k → w(w k· g (sk)k)
⫺ (T(sk+1)k⫺T(sk)k)⫹ ⫺h(sk)k⫺
→ 2 ∂U0 → → 2
||w k|| ||w k|| ||w k||

冉 冊
→ → → →
g (U0)Ng → →
(w Ng−w gNg)·g(U0)Ng ∂
⫹ ⴚ(w Ng⫺w gNg) (Q(sNg+1)NgⴚQ(U0)Ng)
→ → → → 3 ∂U0
||w Ng−w gNg|| ||w Ng−w gNg||

冉 冊
→ → →
g(sNg)Ng → w Ng· g (sNg)Ng ∂Q(U0)Ng
ⴙ ⫺w Ng
→ → 3 ∂U0
||w Ng|| ||w Ng||
→ →
w Ng−w gNg ∂ w Ng ∂T(U0)Ng

⫺ (T(sNg+1)NgⴚT(U0)Ng)ⴚ (A2)
→ → 2 ∂U → 2 ∂U0
||w Ng−w gNg|| 0
||w Ng||

where the supplementary functions are given as follows:

∂Q(s)k

⫽ (A3)
∂J

→ → → 2 → → → → → → 2→
Ng−w gNgm(s)|| −(w Ng−w gNgm(s))·g(s)k)兵||w Ng−w gNgm(s)|| g(s) −((w Ng−w gNgm(s))·
T
((s||w
→ → → → → → → → 2 → → 2 →
Ng−w gNgm(s)) 其+(w Ng−w gNgm(s)) (|| g(s)k|| ||w Ng−w gNgm(s)|| −((w Ng−w gNgm(s))·g(s)k) ))
T T 2
g(s)k)(w
→ 2 → → → 2
→ →
Ng−w gNgm(s)|| −[(w Ng−w gNgm(s))·g(s)k] )
2 3/2
cosh(Q(s)k)(|| g (s)k|| ||w

→ → →
∂T(s)k g (s)k −s(w k−w gkm(s))
T T


⫽ (A4)
→ →
∂J ||w k−w gkm(s)||T(s)k

∂Q(s)k
∂U0
⫽ (A5)

→ 2 → 2 → 2 → 1
k−w gkm(s)|| −(g(s)k·(w k−w gkm(s))) )(||w k−w gkm(s)|| H(s−U0)−(w k−w gkm(s))h(s)k))+
→ → → 2 → → → →
((|| g (s)k|| ||w
2
2 → 2 → → → →
k−w gkm(s)|| −g(s)k·(w k−w gkm(s)))(2||w k−w gkm(s)|| (g (s)k·h(s)k)+2(g(s)k·(w k−w gkm(s)))(w k−w gkm(s))·h(s)k))
→ → → → → → → → → →
(s||w
→ 2 → 2 →
(||g(s)k|| ||w k−w

gkm(s)|| −(g(s)k·(w k−w gkm(s))) ) cosh(Q(s)k)
→ → 2 3/2
Y.T. Chai et al. / Ocean Engineering 29 (2002) 627–649 649

∂T(s)k
⫽ (A6)
∂U0
→ → 2 → → → → → → → →
−2s(H(s−U0)||w k−w gkm(s)|| −(w k−w gkm(s))·h(s)k)+2g(s)k·h(s)k+2(w k−w gkm(s))·g(s)k
→ →
2T(s)k||w k−w gkm(s)||


兵0 0 −(wmx Ng+wcx Ng)其T sⱖsNg+1

h(s)k⫽ 兵wxNg wyNg 0其T sNg+1⬎sⱖU0 (A7)
兵0 0 0其T otherwise

The analytical coefficients given above are derived assuming that the rate of
change of the external drag-loading vector with respect to the anchor tension vector
is small. This assumption is valid in most practical situations except in high drag-
to-weight ratio situations in which case the numerical technique will have to be
applied to account for this effect.

References

Broyden, C.G., 1965. Mathematics of Computation 19.


Chiou, R.B., Leonard, J.W., 1991. Steady response of multi-leg moorings by direct integration. Inter-
national Journal of Offshore and Polar Engineering 1 (3), 176–182.
Huang, S., Vassalos, D., 1993. A semi-analytical treatment of three-dimensional static of marine cables.
Ocean Engineering 20 (4), 409–420.
Leonard, J.W., 1979. Newton–Raphson iterative method applied to circularly towed cable–body system.
Engineering Structure 1 (2), 73–80.
Leonard, J.W., Nath, J.H., 1981. Comparison of finite element and lumped parameter methods for oceanic
cables. Engineering Structures 3 (3), 153–167.
O’Brien, T., 1967. General solution of suspended cable problem. Journal of Structural Division 93 (ST1),
1–26.
Oil Companies International Marine Forum, 1994. Prediction of Wind and Current Loads on VLCCs,
2nd Edition. Witherby and Co. Ltd, London.
Peyrot, A.H., 1980. Marine cable structures. Journal of Structural Division 106 (ST12), 2391–2404.
Peyrot, A.H., Goulios, A.M., 1979. Analysis of cable structures. Computers and Structures 10 (5),
805–813.
Skop, R.A., 1988. Mooring systems: a state-of-the-art review. International Journal of Offshore and Mech-
anics and Arctic Engineering 110 (4), 365–372.
Skop, R.A., O’Hara, G.J., 1970. The method of imaginary reactions: a new technique for analysing struc-
tural cable systems. Marine Technology Society Journal 4 (1), 21.
Skop, R.A., O’Hara, G.J., 1972. A method for the analysis of internally redundant structural cable arrays.
Marine Technology Society Journal 6 (1).
Webster, R.L., 1975. Non-linear static and dynamic response of underwater cable structures using finite
element approach. In: Proceedings of the 7th Offshore Technology Conference, Houston, OTC 2322,
pp. 753–764.

You might also like