See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/262817501

Dynamics of Single Point Mooring Configurations

Conference Paper · September 2007

CITATION READS

1 1,604

1 author:

Thomas Erling Schellin

University of Duisburg-Essen
102 PUBLICATIONS   932 CITATIONS   

SEE PROFILE

Some of the authors of this publication are also working on these related projects:

NEWDRIFT View project

WAVELODS View project

All content following this page was uploaded by Thomas Erling Schellin on 02 December 2015.

The user has requested enhancement of the downloaded file.

 26

OSV Singapore 2007

Jointly organized by Joint Branch of RINA-IMarEST Singapore and CORE.
24-25 September 2007

Dynamics of Single Point Mooring Configurations

Thomas E. Schellin*

*Rule Development, Analysis of Hull Structures & Damages

Germanischer Lloyd, 20459 Hamburg, Germany

ABSTRACT
This paper presents a theoretical-numerical model to investigate the dynamics of single point
moored vessels subject to current, wind, and waves. The time-domain analysis described is based on
an experimentally validated, comprehensive mathematical maneuvering model of ship maneuvering
equations in three degrees of freedom (surge, sway, and yaw) that considers five sets of forces: (1)
nonlinear quasi-steady hydrodynamic response and control forces, (2) linear memory effects due to
radiated waves, (3) nonlinear mooring restoring force characteristics, (4) empirical wind actions, and
(5) first-order wave forces and second-order wave drift forces. The significant nonlinearities inherent
in single point mooring systems, mainly due to hydrodynamic response and control forces as well as
mooring system restoring forces, may lead to multifarious dynamic phenomena, such as self-sustained
oscillations. Under certain conditions, three practical measures, namely, rudder deflection, reverse
propeller action and asymmetric mooring, are shown to stabilize the moored vessel, thus reducing
motion amplitudes and mooring hawser tensions. Computer simulations for a supertanker are
experimentally validated by comparison with model experiments and, as an extension, specifically
focused on the effects of water depth. Effects of different single point mooring configurations on the
dynamic behavior of a supertanker are examined, comprising buoy mooring, articulated tower
mooring, bow turret mooring, and internal turret mooring. In addition, based on the use of an
alternative frequency-domain method, results of a mooring analysis used for design approval are
presented for an offshore coal transshipment terminal that includes single point mooring systems of a
transshipper and two coal barges.

INTRODUCTION
At unprotected offshore locations, tankers and other vessels are increasingly moored using so-
called single point mooring (SPM) systems. The purpose of this paper is to present rational methods
for analyzing SPM systems. Different configurations are in use. A single buoy mooring (SBM), an
internal turret mooring (ITM), and an articulated tower mooring (ATM) are schematically shown for a
tanker in Fig. 1. The obvious advantage of these configurations is that the vessel is free to assume a
favorable alignment to the prevailing current, wind or waves, thereby substantially reducing tensile
forces in the mooring hawser compared to forces possible if the vessel’s heading were constrained.
However, under certain conditions, the vessel may not attain a stable equilibrium even in a seemingly
innocuous steady environment, but indulge in large amplitude low-frequency oscillations of periodic
or aperiodic nature. The problem is of practical importance due to the formidable environmental
implications of an accident. The crucial aspect is the sometimes unpredictable peak forces in the
mooring hawser threatening a possible line break (Sharma et al. 1994).

U:\OSV Singapore 2007\OSV Singapore 2007 - SPM Configurations.doc

 27

During the last decades, a considerable amount of research has been devoted in the development
of theoretical-numerical methods to predict full-scale behavior of an SPM vessel. Significant
nonlinearities inherent in an SPM system, mainly due to hydrodynamic response and control forces as
well as mooring system restoring forces, are
responsible for difficulties in making accurate
predictions of the vessel’s response, displaying many
features typically associated with nonlinear dynamics
and instability. A computational analysis calls for a
reliable mathematical model incorporating sufficiently
accurate descriptions of hydrodynamic and mooring
restoring forces, a realistic representation of
environmental effects (current, wind and waves) as
well as a precise account of all relevant physical
interactions. Apparently, the earliest studies of SPM
ships were carried out by Wichers (1976), Faltinsen et
al. (1979), and Owen and Linfoot (1985). To ensure
economic offshore exploration, SPM systems for large
floating production, storage and offloading (FPSO)
vessels are vital as exemplified by recent
investigations by, for example, Matter et al. (2007),
Munipalli et al. (2007), and Vazquez-Hernandez
(2007).
Single point moorings are used primarily to
position tankers. For tankers, parameters required for
the specification of all forces in the mathematical
model are usually available. Generally, parameters
(coefficients) that depend on hull geometry, propeller,
and rudder are obtained from planar motion tests with
a captive model in a towing tank. The resulting time-
domain simulations must be of sufficient length to
establish reasonable confidence bounds for the
expected maximum response in the storm duration.
Typically, specified storm duration will be simulated
Fig. 1 Schematic of SPM configurations
several times, using statistical fitting techniques to
(Jiang et al. 1995)
establish the expected maximum response. The time-
domain mathematical model is here presented in some detail, including experimental results that
validated the model. Using this model, computer simulations are performed to demonstrate the effects
of different SPM systems on the dynamic behavior of a supertanker.
For offshore type structures and other vessels, on the other hand, hull dependent parameters are
usually not available. In such cases, a rational method documented, for example, by the American
Petroleum Institute (API 1997) is frequently used to analyze, design or evaluate mooring systems. This
method, most conveniently performed in the frequency-domain, provides an alternative approach to
determine the adequacy and safety of mooring systems. It calls for first defining a mean offset of
vessel displacement due to a combination of mean wind, current, and wave drift forces. The maximum
offset is then the mean offset plus appropriately combined wave-frequency and low-frequency vessel
motions, and a probabilistic approach defines the chance of exceeding the combined high- and low-
frequency motions on average of once in the specified storm duration. Results from the use of this
method, performed in the frequency-domain, are presented for the mooring analysis of an offshore
coal transshipment terminal that includes single point mooring systems of a transshipper and two coal
barges.

U:\OSV Singapore 2007\OSV Singapore 2007 - SPM Configurations.doc

 28

MATHEMATICAL MODEL
Two coordinate systems, see Fig. 2,
describe the motion of the vessel single-
point moored in current, wind, and
waves. First, an earthbound coordinate
system P x0 y0 , conveniently centered at
the effective mooring point P, describes
the position of the vessel’s midship point
O and heading angle ψ as well as the
direction of the mooring hawser ψA. It
also defines the the reference directions Fig. 2 Coordinate systems
of current, wind, and waves (ψC,ψW,ψS).
Second, a vessel-bound coordinate system O x y is used to simplify the description of external forces
acting on the vessel. The center of gravity G of the vessel and the attachment point A of the mooring
hawser have fixed coordinates xG ,yG and xA ,yA, respectively, in the vessel-bound coordinate system.
The relation between horizontal vessel velocity components x& 0 , y& 0 and u, v resolved along
earthbound and vessel-bound coordinates, respectively, is contained in the trajectory equations:
x& 0 = u cosψ − v sinψ (1)
y& 0 = v cosψ + u sinψ (2)
ψ& = r (3)
which also define yaw rate r.
The vessel is treated as a transversely symmertic rigid body having three degrees of freedom: surge
sway, and yaw. Effects of heave, pitch, and roll are believed to be small on extremely slow horizontal
motions and hence neglected. The dynamic equations of maneuvering can then be written in standard
form:
(u& − v r − r 2 xG ) m = X (4)
(v& + u r + r& xG ) m = Y (5)
r&I zz + (v& + u r ) xG m = N (6)
where m is vessel mass and Izz its moment of inertia about a vertical axis through O. The net external
time-varying horizontal force components X,Y resolved along axes x,y and their moment N about O
result, in general, from a complex interaction of various physical phenomena. A simple linear
superposition of five effects is considered:
F = { X , Y , N } = FQ + FM + FA + FW + FS
T
(7)
where subscripts Q,M,A,W,S stand for quasi-steady, memory, anchoring, wind, and sea waves,
respectively; superscript T denotes transpose.
The quasi-steady hydrodynamic response and control force couple FQ is calculated according to
the four-quadrant model of Oltmann and Sharma (1984). The force couple elements are synthesized as
follows:
⎧⎪ X HI + X HL − RT + X P + X R ⎫⎪
FQ = ⎨YHI + YHL + YHC + YP + YR ⎬ (8)
⎪⎩ N HI + N HL + N HC + N P + N R ⎪⎭
where subscripts H,P,R stand for system elements hull, propeller, rudder and I,L,C for physical
mechanisms ideal fluid, lifting, cross-flow effects, respectively; the odd term RT simply denotes
ordinary resisitance to pure longitudinal motion. This force couple depends linearly on accelerations
u& , v&, r& and in a highly nonlinear way on velocities u,v,r as well as on control parameters propeller rate
n and rudder angle δ. The explicit formulations are fully documented in the work cited. However, three
features are worth mentioning. First, the four-quadrant model, unlike many others in common use,

U:\OSV Singapore 2007\OSV Singapore 2007 - SPM Configurations.doc

 29

does not break down near speed reversals (u = n = 0) and is, therefore, specially suitable for simulating
slow vessel motions. Second, this model painstakingly accounts for three-way hull-propeller-rudder
interactions. Third, it incorporates simple empirical corrections for the main Reynolds number
associated scale effects on hull resistance and wake with important ramifications for propeller and
rudder forces.
Standard ITTC (1984) corrections to ship resistance are considered without allowance for the
roughness scale effects. To obtain full scale propeller and rudder forces, the wake fraction for the ship
condition is needed. Again, the ITTC (1984) correction formula is used.
A linear response force couple associated with hydrodynamic memory is calculated by means of a
finite state space model fully described by Jiang et al. (1987). The final result can be summerized as
follows:
[ ]
FM = [a (0) − a (∞)]v& + b (0) − b (∞) v + s 0 (9)
s&n−k = sn+1−k − Ak s0 − Bk v (10)
with k = 0,1, · · · , n, s n +1 = 0 and v = (u , v, r ) T . The state vectors sk effectively store memory effects
of motion history in the time domain. Parameter matrices Ak , Bk of an (n + 1) state space model can
be identified by a least squares method to fit theoretically calculated added mass and damping matrices
a (ω ), b (ω ) in the frequency domain. Thus, this formulation accounts for linear memory effects due to
waves radiated from an oscillating hull.

ENVIRONMENTAL EXCITATION
For slow motion of the moored vessel, the main part of hydrodynamic forces is due to the
prevailing current. Therefore, current forces are treated specially. Unlike others, who treat current
forces by means of modular modeling, the vessel’s velocity relative to the ambient water is substituted
into the four-quadrant maneuvering model, thereby accounting for the entire interaction between
vessel motion and current speed. For a steady uniform current of magnitude VC and direction ψC, the
relations between hull velocity components u,v over ground and urel ,vrel relative to ambient water are
u rel = u − VC cos (ψ C − ψ ) (11)
vrel = v − VC sin (ψ C −ψ ) (12)
The wind-generated force couple, due to a wind acting on the above-water parts of the hull and
superstructure, is calculated as usual by means of the following empirical formula:
⎧ 0.5 ρ AVW2 AT C XW ⎫
⎪ ⎪
FW = ⎨ 0.5 ρ AVW2 AL CYW ⎬ (13)
⎪ 0.5 ρ AVW2 AL LC NW ⎪
⎩ ⎭
where ρA is mass density of air, VW is wind speed (accounting in standardized ways for vertical profile
and turbulance level), AL and AT are longitudinal (broadside) and transverse (head-on) projected
above-water areas, L is vessel length between perpendiculars, and CXW,CYW,CNW are ship-type
dependent force and moment coefficients as functions of wind angle of attack (π – ψW – ψ), see
OCIMF (1994).
Wave forces generated on the hull by the action of ambient sea waves are approximated as the
following sum:
FS = FS(1) + FS( 2 ) (14)
where FS(1) is the first-order force and FS( 2 ) the second-order slowly varying drift force. The first-
order force is constructed by superimposing responses to N individual wave components:

ψ S − ψ ) A j exp[ − ik j ( x0 cosψ S + y 0 sinψ S ) + iω j t ]

N
FS(1) = Re ∑ H (ω
j =1
j, (15)

U:\OSV Singapore 2007\OSV Singapore 2007 - SPM Configurations.doc

 30

with wave amplitude Aj and wave number kj corresponding to wave frequency ωj. The hull-form
dependent complex frequency response vector H is also referred to as the linear transfer function of
wave forces. The second-order slowly varying force is assumed proportional to the local wave
envelope profile squared:
FS( 2) = ξ 2 G (ω1 ,ψ S −ψ ) (16)
where G is the hull-form dependent drift force coefficient vector at the mean wave frequency ω1, also
known as the quadratic transfer function of wave forces. The local wave envelope profile ξ is
computed as follows:
ξ = ς 2 +η 2 (17)
with wave elevation ζ and its Hilbert transform η defined as follows:

ς + iη = ∑ A j exp[− ik j (x0 cosψ S + y0 sinψ S ) + iω j t ]

N
(18)
j =1

Numerical values of linear H and quadratic G transfer functions for a given direction of wave
propagation ψS can be obtained effectively using three-dimensional potential boundary element
methods in the frequency domain.

ALTERNATIVE MOORING CONFIGURATIONS

For a given vessel, hull hydrodynamics are assumed to remain unchanged in different mooring
configurations, i.e., by neglecting hydrodynamic interactions between the mooring system and the
hull. This is the case for single buoy mooring (SBM) and articulated tower mooring (ATM)
configurations. For bow turret mooring (BTM) and internal turret mooring (ITM) configurations,
hydrodynamic interactions between turret and hull are assumed small and, therefore, neglected. Thus,
the difference of alternative mooring systems manifests itself only in the mooring retsoring
characteristics.
Three different polynomial functions define the total quasi-static horizontal mooring system
retsoring force FA in the SBM, ATM, and BTM (ITM) configurations. First, a forth-order polynomial
represents the SBM configuration:
1
FA = [1 + sgn (Δ L A )]C 4 (Δ L A ) 4 (19)
2
where
Δ L A = L A − L AU
Here FA arises from the elasticity of the the hawser and the catenary action of the anchored buoy, and
LA is the instantaneous horizontal distance of attachement point A from effective mooring point P,
whereas LAU is the hawser’s reference length corresponding to the unstretched no-load condition.
Second, a second-order polynomial represents the ATM configuration:
(
FA = 0.5 (1 + sgn (Δ L A )) C1 (Δ L A ) + C 2 (Δ L A ) 2 ) (20)
Here FA results from the elasticity of the hawser and the buoyant restoring force of the inclined
articulated tower. Displacement ΔLA has the same meaning as above.
Third, a third-order polinomial represents the BTM and ITM configurations:
2 3
FA = C1 Δ L A + C 2 Δ L A + C3 Δ L A (21)
Here ΔLA is the displacement of turret position A from its no-load reference position.
The mooring restoring forces act horizontally either at attachement point A on the vessel in the
direction from A to effective mooring point P or at instantaneous turret position A on the vessel in the
direction from A to its no-load reference position. The coefficients C1, C2, C3, C4 can usually be
obtained by using the least squares method to fit statically calculated or empirical load-elongation
data.

U:\OSV Singapore 2007\OSV Singapore 2007 - SPM Configurations.doc

 31

The instantaneous horizontal distance of attachement point A from effective mooring point P is
given by the following relationship:
[ ]
1
L A = (x0 + x A cosψ − y A sinψ ) 2 + ( y0 + x A sinψ + y A cosψ ) 2 2 (22)
Finally, the formulation for force couple FA is as follows:
⎧⎪ FA cos (ψ A −ψ ) ⎫⎪
FA = ⎨ FA sin (ψ A −ψ ) ⎬ (23)
⎪⎩ x A FA sin (ψ A −ψ ) − y A FA cos (ψ A −ψ )⎪⎭
where ψA is the direction of horizontal mooring retoring force relative the x-axis in the vessel-bound
coordinate system O x y .

COMPUTATIONAL PROCEDURE
The basic principles underlying the computational procedure can be stated symbolically in a
concise form, representing system dynamics by equations (1-6) and (9-10) in a canonical form as the
following generalized state equation:
x& = f (x , c , t ) (24)
expressing the rate of change of dynamic state vector x = (u, v, r , x0 , y 0 ,ψ ) of order nd = 6 without
T

( )
memory and x = u , v, r , x0 , y 0 ,ψ , s0T , s1T , ⋅ ⋅ ⋅ , s nT T of nd = 6 + 3(n+1) with memory. Vector f is a
function of the state vector x , of a time independent parameter vector c , and of time t, which is the
independent variable.
If the governing state equation does not contain time explicitly, that is, there exists no time-
dependent external excitation, then the system is autonomous; otherwise, the system is
nonautonomous. In the autonomous mode, the vessel is subject to a steady current, a constant wind,
and to wave forces that are reduced to their time-averaged components by disregarding the purely
oscillatory components. In the nonautonomous mode, time-varying wave forces, comprising
oscillatory first-order wave forces and slowly-varying wave drift forces, are included. Due to the
strong nonlinearity of equation (24), there generally exists no analytical solution, and results can only
be obtained numerically.
Vector c in equation (24) comprises all system parameters required for the specification of forces,
including control parameters (rudder angle and propeller rate), operational parameters (attachment
point location and mooring hawser length), and environmental parameters (current speed, wind
velocity, and incident wave identifiers).
Essentially three groups of parameters (coefficients) specify all forces in our mathematical model.
First, parameters are required for the inertial terms and the empirical formulas for mooring restoring
and wind forces; mostly, these comprise vessel main dimensions and certain coefficients. Second,
depending on hull geometry, propeller and rudder, more than 50 parameters are needed to quantify the
forces in the four-quadrant maneuvering model; generally, these are obtained from planar motion tests
with a captive model in a towing tank. Third, parameters must be identified for the memory associated
forces as well as for the first- and second-order wave forces; usually, these are calculated from three-
dimensional boundary element methods.

STABILITY ANALYSIS
In the autonomous mode, the moored vessel will have equilibrium positions x E , which are
mathematically defined as follows:
f (x E , c ) = 0 (25)
Physically, this equation means that the time rates x& of all state variables have to vanish at system
equilibria. The equilibrium positions are determined as iterative solutions of this nonlinear algebraic
equation. Depending on parameter values chosen, there may exist one or more equilibrium positions
with their associated basins (domains of attraction). The boundary between those basins can be

U:\OSV Singapore 2007\OSV Singapore 2007 - SPM Configurations.doc

 32

complicated, sometimes fractal.

To classify the equilibria, stability analyses are performed in the vicinity of each equilibrium.
Therefore, the nonlinear state equation about the equilibrium x E is linearized as follows:
y& = A y (26)
The perturbation vector y = x − x E is then amenable to analytical solutions. The Jacobian matrix is
∂f
A= (x E , c )
∂x
Local stability of any examined state is usually assessed in the sense of Liaponov by solving a
classical eigenvalue problem.
A − Iσ = 0 (27)

where I is the unit matrix having the same dimension as A , and σ denotes the eigenvalues. If A
has no eigenvalues with a zero real part, then the equilibrium position is hyperbolic or non-degenerate.
Otherwise, it is an elliptical or degenerate equilibrium. The classification of hyperbolic equilibrium is
straightforward. If all real parts are negative definite, the equilibrium state is stable, and the
autonomous system has to asymptotically return to it after a sufficiently small, arbitrary initial
disturbance. If one or more real parts are positive definite, the equilibrium state is unstable. Even
though the initial disturbance is arbitrarily small, the system will almost never return to such an
unstable equilibrium. It may asymptotically wander away to a neighboring stable equilibrium, enter a
periodic orbit (limit cycle), get trapped in a quasi-periodic orbit on a torus, or execute chaotic motion.
In a degenerate case the linear stability criterion is not sufficient for classification of equilibrium. Such
elliptical equilibria define a center manifold in state space, and the corresponding parameter values
define a bifurcation point in parameter space.

NUMERICAL SIMULATIONS
The global behavior of the nonlinear differential equation (24) can generally be approximated by
numerical integration. The operation is symbolized as follows:
{x (0) , c } ⇒ x& = f (x , c , t ) ⇒ x (t ) (28)
where initial states x (0 ) and system parameters c are the input, while system response x (t ) is the
output. Strong nonlinearities in the dynamics, specially in the mooring restoring forces, may cause this
differential equation to become numerically stiff. Therefore, numerical computations call for
integration with controlled accuracy.
To interpret the simulated motions, it is useful to know whether the subject system is conservative
or dissipative. The divergence (div) of the nonlinear flow (equation 24) is as follows:
div f ( x , c , t ) = spur A (29)
The trace (spur) of Jacobian matrix A is defined as follows:
nd
spur A = ∑Aj =1
jj (30)

with
∂f
A= (x , c , t )
∂x
If the time averaged mean values of equation (29) are negative, then the system is dissipative or
nonconservative. However, the distinction between a conservative and a dissipative system is not a
trivial numerical task, due to the required long-term predictions of the trajectory. From the physical
point of view the systems dealt with here have both potential and viscous damping terms that cause the
system to be dissipative. The asymptotic behavior of such dissipative systems can be designated as an

U:\OSV Singapore 2007\OSV Singapore 2007 - SPM Configurations.doc

 33

attractor. Typical attractors are a fixed point, a limit cycle, a torus, and a strange (chaotic) attractor.
One characteristic feature of chaos is the extreme sensitivity to small change in initial conditions,
implying long-term unpredictability.

NUMERICAL PREDICTIONS
The two exemplary supertankers TOKIO MARU and ESSO OSAKA were chosen for
computational studies. Their principal particulars are given in Table 1. Specific system parameter
values constituting the four-quadrant maneuvering model of TOKIO MARU in deep water were
documented by Oltmann and Sharma (1984); corresponding parameter values of ESSO OSAKA in
deep and shallow water, by Jiang and Sharma (1993). Tanker ESSO OSAKA meanwhile has been
recommended by the ITTC as a standard ship for comparative maneuvering studies of all kinds. The
mooring restoring coefficients are quantified as C1 = 176.69 kN/m, C2 = –2.44 kN/m² and C3 =
0.11 kN/m3 for BTM or ITM, as C1 = 15.92 kN/m and C2 = 0.68 kN/m² for ATM, as C4 =
0.0113 kN/m4 (TOKIO MARU) and C4 = 0.025 kN/m4 (ESSO OSAKA) for SBM. The corresponding
horizontal mooring hawser forces are graphically shown in Fig. 3.
Table 1 Principal particulars of subject tankers
TOKIO MARU ESSO OSAKA
Length between perpendiculars 290.0 m 325.0 m
Length at waterline 296.4 m 335.0 m
Beam 47.5 m 53.0 m
Draft 16.1 m 21.8 m
Block coefficient 0.81 0.83
Number of propellers 1 1
Screw sense righthanded righthanded
Number of rudders 1 1
Rudder area 73.5 m² 124.7 m²
Seven environmental parameters
were selected as varying system
parameters: current speed (VC) and
direction (ψC), wind speed (VW) and
direction (ψW), significant wave
height (H1/3), mean wave period (T1)
and wave direction (ψS); two control
parameters: rudder angle (δ) and
propeller rate (n); and two
operational parameters: mooring
hawser attachment location (xA, yA) Fig. 3 Horizontal restoring forces of four mooring
and hawser length (LAU). configurations (Sharma et al. 1994)
Hopf Bifurcations
For the two exemplary supertankers numerous stability analyses were performed in parameter
space and carried out as corresponding time domain simulations in state space, intending to cover a
wide practical range of parameter values. For the tankers subject to a steady current without wind and
waves under single buoy mooring configurations, linearized stability analyses demonstrated that
dynamic (Hopf) bifurcations occur. Mathematically, this means that there is a complex eigenvalue pair
whose real part crosses the stability limit Re[σ] = 0 in parameter space. As current speed is increased
(around VC = 0.45 m/s), the real part of the critical eigenvalue of the Jacobian matrix crosses the
stability limit, indicating transition from stable to unstable equilibrium. The global behavior near this
Hopf bifurcation was investigated by performing two numerical simulations having identical fixed
parameter values except for current speed. At a current speed of VC = 0.2 m/s (before Hopf
bifurcation), the tanker asymptotically returned to its stable equilibrium after initial disturbance. Such

U:\OSV Singapore 2007\OSV Singapore 2007 - SPM Configurations.doc

 34

an attractor is considered a fixed point. As current speed is increased to VC = 0.5 m/s (just after Hopf
bifurcation), even though the initial disturbance is arbitrarily small, the tanker never returned to its
unstable equilibrium but asymptotically entered a limit cycle. The two simulated trajectories of the
SBM tanker ESSO OSAKA subject to current, showing the transition from a fixed point to a limit
cycle, see Sharma et al. (1994).
If the SBM tankers are subjected to the
additional effects of wind and waves,
subcritical Hopf bifurcations occur in
certain parameter subspaces. A critical
feature of subcritical Hopf bifurcations is
the coexistence of a fixed point and a limit
cycle of system behavior. Simulated
trajectories of the SBM tanker TOKYO
MARU subjected to wind and waves
directed opposite to current were carried
out with the following parameter values:
VC = 1.5 m/s, LAU = 75 m, xA = 145 m, yA =
0, n = 0, δ = 2.0°, H1/3 = 2 m, T1 = 10 s,
ship condition. They show the coexistence
of two attractors before the subcritical
Hopf bifurcation (Sharma et al. 1994). For
identical parameters values, depending on
initial conditions chosen, the tanker
asymptotically returns to either a fixed
point or it enters a stable limit cycle.
Global Bifurcations and Chaotic Motion
A common feature of dissipative
systems is that global behavior changes
qualitatively if one parameter value
changes. This phenomenon is shown in Fig. 4 Qualitatively different attractors of SBM tanker
Fig. 4, displaying the rich variety of ESSO OSAKA subject to current in shallow water
attractors attainable by the tanker ESSO (Sharma et al. 1994)
OSAKA in a steady current. Results
shown here are for the ship condition in shallow water with the following parameter values: LAU =
78 m, xA = 162.5 m, yA = 0, n = 0, δ = 0. As current speed increases, the global behavior is first
characterized by a fixed point (a), then a limit cycle of period close to the linearized eigenperiod (b),
next a limit cycle with approximately twice the linearized eigenperiod (c), followed by chaos (d),
reverting surprisingly to a shoe-shaped limit cycle of single eigenperiod (e), and moving on to an oval
limit cycle (f). At the local level, after the supercritical Hopf bifurcation, all cases differ only
quantitatively, i.e., they are all dynamically unstable. The possible pathway from a limit cycle (case b)
to chaotic motion (case d) can be understood here as a sequence of global bifurcations.
Influence of Alternative Mooring Configurations
One of the most important nonlinearities in the dynamic system arises from the mooring restoring
characteristics. They are totally different in alternative mooring configurations; see Jiang et al. (1995).
To demonstrate the influence of mooring restoring force characteristics on stability of the SPM tanker,
linearized stability analyses were performed of tanker TOKYO MARU subject to steady current for
identical conditions, but under buoy mooring (SBM) and articulated tower mooring (ATM)
configurations. The stability analysis revealed that increasing current speed is destabilizing and
increasing rudder deflection is stabilizing. But the primary effect here is that restoring force
characteristics resulting from the different mooring configurations have no remarkable influence on

U:\OSV Singapore 2007\OSV Singapore 2007 - SPM Configurations.doc

 35

Hopf bifurcations. To investigate effects of changing the mooring restoring force characteristics on
global behavior, two simulations were carried out of tanker TOKYO MARU subject to steady current
for identical conditions, and again under buoy mooring (SBM) and articulated tower mooring (ATM)
configurations. These simulations revealed that the
mooring restoring force characteristics affect the
form of the asymptotic motion trajectories as well
as their amplitudes (Jiang et al. 1995).
Influence of Water Depth
Hydrodynamic response and control forces are
the other most important nonlinearity in the
system. They vary strongly with water depth. The
influence of water depth on system behavior is
manifested by two simulations shown in Fig. 5.
The tanker ESSO OSAKA was moored in deep
(solid lines) and shallow (dashed lines) water
under otherwise identical conditions: VC = 1.5 m/s,
LAU = 78 m, xA = 162.5 m, yA = 0, n = 0, δ = 0,
model condition. Simulated time histories show
that hawser tension peaks are an order of
magnitude higher in shallow water than in deep
water. Furthermore, the asymptotic trajectories are Fig. 5 Simulated time histories and trajectories of
qualitatively different. Whereas the tanker enters a SBM tanker ESSO OSAKA subject to steady
current in deep and shallow water
periodic limit cycle in deep water, it executes a
(Sharma et al. 1994)
chaotic attractor in shallow water with larger
excursions.
Stabilizing Measures
The practical consequences of self-sustained
oscillations or even chaotic motions of the SPM
tanker are that they cause large motion amplitudes
and high line tensions peaks. The large motion
amplitudes impair safe operations of the tanker,
and high hawser peaks threaten a possible hawser
break leading to a tanker accident. Simple
practical measures have been found to stabilize the
equilibrium states, thereby reducing motion
amplitudes and mooring hawser tensions. They are
static rudder deflection, reverse propeller rate, and
asymmetric fairlead location or their suitable
combinations, see Jiang et al. (1987) and Sharma Fig. 6 Simulated time histories and trajectories of
et al. (1988). Figure 6 confirms the stabilizing SBM tanker ESSO OSAKA subject to current with
effect of static rudder deflection for the SBM and without rudder application (Sharma et al. 1994)
tanker ESSO OSAKA in ship condition in shallow
water. Simulated time histories show strongly reduced hawser tension peaks when rudder is applied
(dashed line). Trajectories asymptotically lead to a fixed point with rudder hard to starboard
(δ = –35°), but to a chaotic attractor with rudder amidships (solid line).

EXPERIMENTAL VALIDATIONS
To validate the numerical predictions, model experiments were conducted in the main towing tank
(200 m x 9.8 m x 1 m) of the shallow water research facility DST in Duisburg (Development Centre

U:\OSV Singapore 2007\OSV Singapore 2007 - SPM Configurations.doc

 36

for Ship Technology and Transport Systems, formerly VBD), which has the feature of controllable
water depth from zero to one meter. A 1:65 scale model of tanker ESSO OSAKA was used. The main
purpose of model experiments was to physically simulate and record the dynamic behavior of an SPM
tanker in a steady current. Specifically, the
following quantities were measured as functions
of time: horizontal motions of the model in three
degrees of freedom (surge, sway, and yaw) and
mooring hawser tension as well as elongation.
Froude dynamic similarity was assumed, i.e., no
attempt was made to compensate for the
difference in Reynolds number between model
and full-scale. Thus all results of model
experiments correspond to numerical simulations
for the model condition. Current speed was
simulated by towing the model through still
water. The nonlinear load-elongation
characteristics (see Fig. 3) were dynamically
modeled by employing, in series, a cascade of
several linear springs of successively increasing
stiffness and individually bounded extension. For
details of model experiments and measured
results, see Jiang and Sharma (1992).
Self-Sustained Oscillations
A calculated time history (dashed lines) is Fig. 7 Measured and calculated time histories of SBM
compared to a corresponding measured time tanker ESSO OSAKA in shallow water
history (solid lines) in Fig. 7 for the SBM tanker (Jiang and Sharma 1992)
subject to a steady current in shallow water. The
parameter values are as follows: VC = 2.0 m/s, LAU = 78 m, xA = 162.5 m, yA = 0, n = 0, δ = 0. For this
locally dynamically unstable case identified by stability analysis, the corresponding horizontal motions
from numerical simulation as well as from model measurement are characterized by slow self-
sustained oscillations (fish-tailing motions), which are often observed in practice when tankers are
moored to SPM systems. Furthermore, these slow motions cause high loads in the mooring hawser as
seen in the time histories of hawser tension. The agreement between calculation and measurement is
remarkable for motions as well as for the high peaks in hawser tension. Similar agreements were also
achieved for other compared cases, thus experimentally validating the mathematical model.
Influence of Water Depth
To experimentally demonstrate the influence of water depth, model tests were conducted for the
SBM tanker subject to a steady current in deep and shallow water under otherwise identical parameter
values, but different initial conditions. The corresponding time histories for deep and shallow water are
documented by Sharma et al. (1994). Both cases reflect dynamically unstable equilibria (which can be
numerically verified by stability analysis); however, motion amplitudes are much larger and hawser
tension several times higher in shallow water than in deep water.
Stabilizing Measures
To validate the effectiveness of the numerically found stability measures, measured time histories
with rudder deflection were compared with those with rudder amidships. Results showed that when
rudder is applied to a static value of δ = –35°, self-oscillations of tanker motion and high peaks in
hawser tension both disappear after the initial disturbance. Reverse propeller rate, as another
stabilizing measure, works in the same manner.

U:\OSV Singapore 2007\OSV Singapore 2007 - SPM Configurations.doc

 37

SPM MOORINGS FOR AN OFFSHORE TRANSSHIPMENT TERMINAL

In 2002, Oldendorff Carriers (2006)
installed the world’s largest offshore coal
transshipment terminal in the Bay of
Iskenderun, Turkey. Around 20 capesize
shipments with about 3.2 million tons arrive
annually from Columbia and South Africa to
feed the coal to a stockpile located adjacent to
a recently commissioned power station. The
transshipper ISKEN, a twin-hulled, non-
propelled floating cargo terminal, and two
gravity type selfunloading hopper barges of
118 m length and 27 m breadth are part of the
terminal. While one barge is loaded by the
transshipper, the other barge is busy with
Fig. 8 The transshipper loading coal from a bulker
unloading at the shoreside pier. The water
into a hopper barge
depth at the pier is only 6.0 m, so the barges
are designed to carry about 10,000 t on a shallow draft of 4.95 m. With a length of 107 m, a breadth of
44 m, and a total height of 53 m, the transshipper is an impressive steel structure displacing 5700 t.
For weathering severe storms, the transshipper and
the two hopper barges can be towed to three single point
mooring buoys. Each buoy is moored the sea bed by six
catenary anchor leg mooring (CALM) chains. A
mooring hawser holds the vessels captive to a turntable
mounted on top of the buoys by means of a slewing
bearing. This bearing allows the turntable to freely
weathervane, and the moored vessel can take up the
most favorable alignment of least resistance to the
prevailing current, wind and waves.
The mooring analysis was based on three-hour
storm duration. The Deutsche Wetterdienst (DWD),
Fig. 9 One of three mooring buoys under
Hamburg, furnished wave statistics for the site. A
construction
constant one-minute average wind speed of VW = 34.1
m/s, a current speed of VC = 0.75 m/s, and a seaway with a significant wave height of H1/3 = 6.5 m and
a mean wave period of T1 = 6.9 s characterized the environmental conditions.
The severity of these conditions relative to the shallow nominal water depth of 42 m called for a
mooring system flexible enough to tolerate large horizontal motions of the moored vessels. The
CALM buoy by itself did not supply sufficient horizontal flexibility. A long stretchable hawser was
required to produce the needed overall flexibility of the system. The mooring systems comprise six 76
mm diameter Grade 3 stud link chain cables of 275 m length and six Stevpris anchors of adequate
holding power. Each mooring chain, pretensioned to 834 kN, extends from the 7.0 m diameter
mooring buoys in a spread pattern at 30° intervals around each buoy. Figure 9 shows one of the
mooring buoys under construction.
A quasi-static analysis method was performed to evaluate the performance of the SBM systems,
following the recommendations for design and analysis of floating production systems (API 2001) and
station keeping systems for floating structures (API 1997). The effects of line dynamics were
accommodated through the use of a relatively conservative safety factor. To account for possible
unsteady yaw motions, the moored vessels were analyzed under a 15° angle of attack, i.e., ψA = 15°
and ψW = ψC = ψS = 180°.
Wind, current, and wave forces were assumed to act collinearly on the moored vessels, subjecting

U:\OSV Singapore 2007\OSV Singapore 2007 - SPM Configurations.doc

 38

same to a mean steady-state environmental force in the horizontal direction. These mean forces cause
the vessels to be displaced from their initial no-load equilibrium position to a new equilibrium
position, referred to the mean horizontal offset. The wave forces cause the vessels to oscillate about
this mean offset. The vessels’ maximum horizontal offset is thus the sum of the steady-state mean
offset and the wave-induced horizontal displacement. The wave-induced displacement comprises first-
order (high-frequency) motion response at wave frequencies as well as second-order (low-frequency)
slowly varying wave drift motion response. Tensile forces in the mooring hawser and the anchor
chains are then a function of the maximum horizontal offsets. These forces were evaluated quasi-
statically, with appropriate safety factors accounting for dynamic effects.
Standard spectral techniques determined oscillatory first-order wave-induced motions, the mean
horizontal wave drift forces, as well as the associated drift displacements of the moored vessels.
Dedicated computer codes yielded transfer
functions of first-order motions (Papanikolaou
and Schellin 1992), second-order (quadratic)
drift force coefficients (Clauss et al. 1982), and
drift damping coefficients (Schellin and Kirsch
1989). The JONSWAP spectrum with a
peakedness parameter of 3.3 defined the wave
spectral energy density of the seaway. A cosine Fig. 10 Schematic of the SBM systems
squared spreading function described of wave
energy distribution about the principal wave direction.
Another computer code (Schellin and Scharrer 1981) evaluated restoring forces of the SBM
systems, based on catenary equations modified for elastic stretch and bottom friction of the anchor
chains as well as elastic stretch of the mooring hawsers. The mooring systems as analyzed are
schematically depicted in Fig. 10.
Table 3 summarizes the resulting mean environmental forces acting on the moored vessels. For the
transshipper and the barges, laden as well as in ballast, the steady sate force due to wind turned out to
be the dominant part of the total mean environmental force acting on the moored vessels.
Table 3 Mean environmental forces
Mean force [kN] Transshipper Barge (laden) Barge (in ballast)
Total mean environmental force 2496 1038 1135
Mean wind force 2302 777 984
Mean current force 21 36 27
Mean wave drift force 173 225 124

The expected maximum horizontal offset was obtained by adding to the mean offset a combined
value of high- and low-frequency horizontal displacements. This combined value was the sum of the
maximum value of the dominant component and the significant value of the other component, where
the dominant component was that with the highest value. For all three SPM systems investigated,
high-frequency vessel motions were the dominant component as they always exceeded low-frequency
motions. Maximum horizontal offset xmax was thus obtained as follows:
xmax = x E + x HF max + x LF sig (31)
where x E is the mean offset, x HF max is the statistical maximum amplitude of the high-frequency
horizontal motion, and x LF sig is the significant amplitude of the low-frequency horizontal motion. The
statistical maximum and significant values were calculated as follows:
x HF max = x HF sig [ 0.5 ln (T ⋅ f 0 )] 1 / 2 and x LF sig = 2σ LF (32)
where T is the time in seconds of the three-hour storm duration, f 0 is the average up-crossing
frequency of the high-frequency motion, and σ LF is the standard deviation of the low-frequency
motion.

U:\OSV Singapore 2007\OSV Singapore 2007 - SPM Configurations.doc

 39

The method used to yield maximum tension MBL = 4218

in the most loaded anchor chain and maximum
4000
force in the mooring hawser is illustrated in Fig.

F O R C E [k N ]
11, here for the transshipper single-point
moored using a 270 m long hawser with an THAWSER = 2926
elastic coefficient of 100 kN/m. (In this figure, 3000

FE stands for total mean environmental force, FE = 2496

TCHAIN for chain tension, THAWSER for hawser

TCHAIN = 2274
force, MBL for minimum breaking strength, xE 2000

for mean offset, and xMax for maximum offset.)

For the transshipper, a laden barge, and a barge
in ballast, Table 4 summarizes the resulting 1000

maxima of chain tension, hawser force,

horizontal offset, and hawser stretch. Safety
XE = 28.19 XMax = 32.70
factors against minimum break load for the 0
most loaded anchor chain and the hawser are 0 10 20 30 40 50 60 70 80

given as well. According to design conditions OFFSET [m]

for a severe storm, the quasi-static method MOST LOADED CHAIN HAWSER FORCE

applied here required a safety factor of 1.8

Fig.11 Chain tension and hawser force
applied against the minimum rated breaking
for the moored transshipper
strength of the anchor chains and the hawser
(Germanischer Lloyd 2007).

Table 4 Maximum chain tension, hawser force, horizontal offset, and hawser
stretch for the transshipper, a laden barge, and a barge in ballast

Maxima Transshipper Barge (laden) Barge (in ballast)

Chain tension (SF) 2274 kN (1.9) 1388 kN (3.0) 1556 kN (2.7)
Hawser force 2926 kN 1313 kN 1632 kN
Horiz. Offset 32.3 m 15.4 m 18.9 m
Hawser stretch 29.3 m 13.1 m 16.3 m

CONCLUSIONS
An experimentally validated, comprehensive mathematical model is presented and applied to
investigate the dynamics of different configurations of single point moored vessels. Local linearized
stability analyses and nonlinear numerical simulations of an SPM tanker reveal that multifarious
dynamic phenomena may occur, leading to large motion amplitudes and high line tension peaks.
Simple practical measures, e.g., static rudder deflection and reverse propeller rate can be applied to
stabilize the tanker’s equilibrium states, thereby reducing motion amplitudes and mooring hawser
tensions.
Restoring force characteristics resulting from alternative mooring configurations have no
noticeable influence on local stability, but they strongly affect the global behavior. Water depth is a
relevant parameter; motion amplitudes and hawser tension peaks are higher in shallow than in deep
water.
For SPM systems where some of the parameters required for the application of the presented
mathematical model are unavailable, the rational analysis method recommended by API (1997),
combined with an understanding of the environment, the characteristics of the vessel being moored,
and other relevant factors, can be used to evaluate single point mooring systems.

U:\OSV Singapore 2007\OSV Singapore 2007 - SPM Configurations.doc

 40

REFERENCES
API (2001). “Recommended Practice for Planning, Designing, and Constructing Floating Production Systems,”
American Petroleum Institute, RP 2FPS, Washington, D.C.
API (1997). “Recommended Practice for Design and Analysis of Station Keeping Systems for Floating
Structures,” American Petroleum Institute, RP 2SK, Washington, D.C.
Clauss, G., Sükan, M., and Schellin, T.E. (1982). “Drift Forces on Compact Offshore Structures in Regular and
Irregular Waves,” J. Applied Ocean Research, Vol. 4, pp. 208-218.
Faltinsen, O.M., Kjaerland, O., Liapis, N., and Walderhaug, H. (1979). “Hydrodynamic Analysis of Tankers at
Single-Point Mooring Systems,” Proc. 2nd Int. Conf. on Behavior of Offshore Structures, London, Paper No.
BOSS 59.
Germanischer Lloyd (2007). Rules for Classification and Construction IV, Industrial Installations, 6 Offshore
Technology, Hamburg.
ITTC (1984). Proc. of the 17th Int. Towing Tank Conf., Göteborg.
Jiang, T., Schellin, T.E., and Sharma, S.D. (1995). “Horizontal Motions of an SPM Tanker Under Alternative
Mooring Configurations,” J. Offshore Mechanics and Arctic Engineering, Vol. 117, pp. 223-231.
Jiang, T. and Sharma, S.D. (1992). “Investigation of Horizontal Motions of an SPM Tanker in Shallow Water
Through Computation and Model Experiments,” Proc. 19th Symp. on Naval Hydrodynamics, Seoul, National
Academy Press, Washington, D.C., pp. 405-424.
Jiang, T., Schellin, T.E., and Sharma, S.D. (1987). “Maneuvering Simulation of a Tanker Moored in a Steady
Current Including Hydrodynamic Memory Effects and Stability Analysis,” Proc. Int. Conf. on Ship
Manoeuvrability, RINA, London, Vol. 1, Paper No. 25.
Matter, G.B., Sphaier, S.H., and Sales Jr., J.S. (2007). “Definition of the Best Position for the Turret in an FPSO
Based on a Hydrodynamic-Structural Analysis,” Proc. 26th Int. Conf. on Offshore Mechanics and Arctic
Engineering, ASME, San Diego, Paper No. OMAE 2007-29474.
Munipalli, J., Pistani, F., Thiagarajan, K.P., Winsor, F., and Colbourne, B. (2007). “Weathervaning Instabilities
of an FPSO in Regular Waves and Consequence on Response Amplitude Operators,” Proc. 26th Int. Conf.
on Offshore Mechanics and Arctic Engineering, ASME, San Diego, Paper No. OMAE 2007-29359.
OCIMF (1994). Prediction of Wind and Current Loads on VLCCs. Oil Companies Int. Marine Forum, London,
Witherby & Co. Ltd.
Oldendorff Carriers (2006). “Transshipper ISKEN in Turkey – 40,000 tpd,” Update on our Coal Transshipment
Activities, Lübeck, Germany.
Owen, D.G. and Linfoot, B.T. (1985). “Theoretical Analysis of Single Point Mooring Behavior,” Trans.
SNAME, Vol. 85, pp. 315-324.
Papanikolaou, A.D. and Schellin, T.E. (1992). “A Three-Dimensional Panel Method for Motions and Loads for
Ships with Forward Speed,” J. Ship Technology Research, Vol. 39, pp. 147-156.
Schellin, T.E. and Kirsch, A. (1989). “Low-Frequency Damping of a Moored Semisubmersible Obtained from
Simulated Extinction Tests and Mean Drift Forces,” J. Applied Ocean Research, Vol. 4, pp. 202-211.
Schellin, T.E. and Scharrer, M. (1981). “Design Principles to Select Effective Mooring Systems, Hansa, Vol.
118, No. 6, pp. 432-438 (in German).
Sharma, S.D., Jiang, T., and Schellin, T.E. (1994). “Nonlinear Dynamics and Instability of SPM Tankers,” Fluid
Structure Interaction in Offshore Engineering, S.K. Chakrabarti (ed.), Computational Mechanics
Publications, Ashurst, UK, pp. 85-123.
Sharma, S.D., Jiang, T., and Schellin, T.E. (1988). “Dynamic Instability and Chaotic Motions of a Single-Point-
Moored Tanker, Proc. 17th Symp. on Naval Hydrodynamics, The Hague, National Academy Press,
Washington, D.C., pp. 543-563.
Vazquez-Hernandez, A.O. (2007). “FPSO Conceptual Design System Tools Considering Hurricane Data Base
and Production Requirements,” Proc. 26th Int. Conf. on Offshore Mechanics and Arctic Engineering,
ASME, San Diego, Paper No. OMAE 2007-29102.
Wichers, J.E.W. (1976). “On the Slow Motion of Tankers Moored to Single-Point Mooring Systems,” Proc.
Offshore Technology Conference, Houston, Paper No. OTC 2548.

U:\OSV Singapore 2007\OSV Singapore 2007 - SPM Configurations.doc

View publication stats