Global Performance Analysis of Deepwater Floating Structures
Global Performance Analysis of Deepwater Floating Structures
Global Performance Analysis of Deepwater Floating Structures
DNV-RP-F205
The electronic pdf version of this document found through http://www.dnv.com is the officially binding version
© Det Norske Veritas
If any person suffers loss or damage which is proved to have been caused by any negligent act or omission of Det Norske Veritas, then Det Norske Veritas shall pay compensation to such person
for his proved direct loss or damage. However, the compensation shall not exceed an amount equal to ten times the fee charged for the service in question, provided that the maximum compen-
sation shall never exceed USD 2 million.
In this provision "Det Norske Veritas" shall mean the Foundation Det Norske Veritas as well as all its subsidiaries, directors, officers, employees, agents and any other acting on behalf of Det
Norske Veritas.
Recommended Practice DNV-RP-F205, October 2010
Changes – Page 3
CHANGES
• General • Main changes
As of October 2010 all DNV service documents are primarily Since the previous edition (October 2004), this document has
published electronically. been amended, most recently in April 2009. All changes have
been incorporated and a new date (October 2010) has been
In order to ensure a practical transition from the “print” scheme given as explained under “General”.
to the “electronic” scheme, all documents having incorporated
amendments and corrections more recent than the date of the
latest printed issue, have been given the date October 2010.
An overview of DNV service documents, their update status
and historical “amendments and corrections” may be found
through http://www.dnv.com/resources/rules_standards/.
CONTENTS
mooring and riser system be relocated, but is generally positioned at the same location
3) Seafloor friction (if mooring lines and/or risers have bot- for a prolonged period of time. The unit normally consists of a
tom contact) ship hull, with turret, and production and drilling equipment on
Damping: deck. For FPSOs, due to their large superstructures and their
active or passive weather-vaning ability, wind forces are often
4) Damping from mooring and riser system due to dynamics, dominant relative to current forces. FPSOs normally experi-
current, etc. ence significant LF response in the horizontal plane. They may
5) Friction forces due to hull/riser contact. be particularly sensitive to surge excitation due to the low vis-
cous hull damping. This sensitivity is reduced with increasing
Inertia: water depth since the damping contributions from mooring
lines and risers increase.
6) Additional inertia forces due to the mooring and riser sys-
tem. FPSOs are flexible with respect to selection of deep water
mooring systems. For catenary mooring systems, the WF
In a traditional de-coupled analysis, item 1) can be accurately motions can introduce dynamic mooring forces, which tend to
accounted for. Items 2), 4) and 6) may be approximated. Gen- increase in deep water due to larger transverse drag forces.
erally, items 3) and 5) cannot be accounted for. A coupled Taut mooring systems are not subjected to the same level of
analysis as described previously can include consistent treat- transverse motions, thus acting more quasi-statically. Dynamic
ment of all these effects. forces will tend to decrease with increasing water depth for
2.1.3 De-coupled analysis such systems, since the elastic length of the mooring lines
increases. Fishtailing is the unstable coupled yaw and sway
In a de-coupled analysis the equations of the rigid body floater motions excited by wind and current. It is associated with the
motions are solved in time domain, but the effects of the moor- horizontal stiffness of the mooring system. For riser systems,
ing and riser system are included quasi-statically using non- flexible risers and compliant metallic risers are usually applied
linear springs, i.e. quasi-static restoring force characteristics. due to the significant WF motions.
All other coupling effects, e.g. contributions from damping
and current loading on the slender structures, need to be given FPSOs may have one or several moonpools, and the water
as input to the analysis based on a separate assessment. motion in the moonpool can influence the vessel motions. Vis-
cous damping has a strong influence on this water motion.
2.1.4 Coupled analysis Slamming and green water on deck are other non-linear effects
that may influence FPSO response in rough weather.
In a coupled analysis the complete system of equations
accounting for the rigid body model of the floater as well as the Combination of wind generated waves and swell with different
slender body model for the risers and mooring lines are solved headings are a challenge and must be taken into consideration.
simultaneously using a non-linear time domain approach for This applies to turret moored vessels as well as vessels with
dynamic analyses. Dynamic equilibrium is obtained at each spread mooring. A critical condition is the combination of
time step ensuring consistent treatment of the floater/slender head sea and beam swell. Significant roll accelerations may
structure coupling effects. The coupling effects are automati- occur and thus have impact on topside structure and equip-
cally included in the analysis scheme. ment, riser system and mooring system etc.
2.2 Main characteristics of floaters Selection of proper roll damping is important in the prediction
of FPSO responses.
A common feature of all types of floaters is that they utilise
excess buoyancy to support deck payload and provide slender Floating systems involving multiple floaters have been
structure tensions. Depending on the area and the sea state, designed and installed. A typical field architecture may con-
ocean waves contain 1st harmonic wave energy in the period sist of a spread-moored FPSO and a dry tree unit (DTU), e.g.
range of 5 - 25 s. For a floating unit the natural periods of Spar, TLP or barge, connected by fluid transfer lines (FTLs).
motions are key features and in many ways reflect the design The offloading system (e.g. CALM buoy) can be a few kilome-
philosophy. Typical motion natural periods of different float- tres away from the FPSO and connected to the FPSO through
ers are presented in Table 2-1. oil offloading lines (OOLs). These complex multi-floater sys-
tems bring additional challenges to both model testing and
Table 2-1 Typical natural periods of deep water floaters
numerical analyses. From the analysis point of view, the fol-
lowing issues are of importance:
Natural periods (seconds)
Floater — consistency in phasing of waves and loads
FPSO DDF TLP Semi — wind-generated waves, swell and current with different
Mode
Surge > 100 > 100 > 100 > 100 headings
— additional coupling effects due to FTLs and OOLs
Sway > 100 > 100 > 100 > 100 — possible hydrodynamic interactions between floaters.
Heave 5 – 12 20 – 35 <5 20 – 50
If the two floaters (FPSO and DTU) are close enough to each
Roll 5 – 30 50 – 90 <5 30 – 60 other, hydrodynamic interactions related to wave effects can
Pitch 5 – 12 50 – 90 <5 30 – 60 be of importance. This requires a hydrodynamic analysis of the
two floaters as an integrated system with 12 degrees of free-
Yaw > 100 > 100 > 100 > 100
dom using diffraction/radiation theory.
A common characteristic of all floater types is that they are All the above effects may be included in a computer simulation
“soft” in the horizontal plane, with surge, sway and yaw peri- program designed for multiple floaters and their associated
ods generally longer than 100s. The fundamental differences slender structures.
among the floaters are related to their motions in the vertical
plane, i.e. heave, roll and pitch. The floater motions in the ver- 2.2.2 TLP response characteristics
tical plane are decisive for the choice of riser and mooring sys- A TLP differs fundamentally from other floater concepts in the
tems. sense that it is the tendon stiffness rather than the waterplane
stiffness that governs the vertical motions. The TLP is a soft
2.2.1 FPSO response characteristics spring in surge, sway and yaw motions, but stiff in heave, roll
A floating production storage and offloading unit, FPSO, can and pitch motions.
A TLP generally experiences WF motions in the horizontal The semi-submersible is very sensitive to weight changes; i.e.
plane that are of the same order of magnitude as those of a it has low flexibility with respect to deck load and oil storage.
semi-submersible of comparable size. In the vertical plane,
however, the TLP will behave more like a fixed structure with Compared to ship-shaped floaters, the current forces will be
practically no WF motion response. WF forces are directly larger on semi-submersibles due to the bluff shapes of their
counteracted by the tendon stiffness forces. underwater columns and pontoons. Wind loads will still dom-
inate the mean forces, except in calm areas with strong cur-
Higher order sum-frequency wave forces may introduce rents.
springing or/and ringing responses in the vertical modes.
These effects may give significant contributions to the tether The semi-submersible is characterized by having free modes
responses. of motion only, which means that all natural periods are above
the range of natural wave periods, see Table 2-1. Despite this
Set-down is the kinematic coupling between the horizontal fact, the wave frequency motions are not insignificant, espe-
surge/sway motions and the vertical heave motions. Set-down cially in extreme conditions, as indicated in Figure 2-1.
is important in the calculation of airgap, tether forces and riser
system responses such as stroke.
The TLP riser system typically consists of top tensioned risers,
flexible risers or compliant metallic risers such as steel cate-
nary risers.
2.2.3 DDF response characteristics
A Deep Draught Floater (DDF) is characterised by small
heave motions. An example of a DDF is a Spar platform. The
main hull of a Spar is a cylinder with a central moonpool for a
riser system in tension. The hard tank provides buoyancy and
the part below may consist of a shell structure (Classic Spar),
or a truss structure (Truss Spar) with a soft tank at the keel and
added mass/damping plates in between. The Spar has a large
area exposed to current forces, which is usually the dominant
environmental load. LF vortex induced motions (VIM) may
increase the effective drag leading to even higher mean current
forces. By adding strakes on the Spar hull, the vortex induced
cross-flow oscillation can be reduced by considerable amount
However, the strakes will increase the added mass and the drag
forces on the Spar.
The small heave motions of a DDF allows the use of rigid top-
tensioned vertical risers. The riser tension is normally provided
by either air cans attached to the upper part of the risers, or by
tensioners integrated to the hull. Spars using air can supported
risers are characterized by having free modes of motion only.
Their heave natural period is usually above the range of wave
periods. Spars with tensioner supported risers experience Figure 2-1
greater coupling in heave, since the heave restoring and heave Heave transfer functions for different floaters and storm wave
eigenperiod are influenced by the riser system. This means that spectrum
a heave damping assessment is crucial for the prediction of the
Spar heave response.
Current fluctuations may induce significant excitation forces Large semi-submersibles with displacement of 100000 tonnes
on a DDF. Depth correlation is a central issue when determin- or more are generally less sensitive to WF action. LF responses
ing the level of such excitation. may be more dominating in roll and pitch motions.
Air-gap and moonpool effects should be considered for Spar Wave impact underneath the deck due to insufficient air-gap
analysis and design. may influence the global motions and local structural
responses for semi-submersibles.
Due to low WF motions, a DDF is generally not subjected to
large dynamic mooring line forces. This has to be evaluated in Catenary moored semi-submersibles may experience signifi-
relation to the actual location of the fairleads and the increase cant dynamic mooring forces due to WF responses similar to
in horizontal WF motion towards the waterline. those of a FPSO.
2.2.4 Semi-submersible response characteristics 2.3 Main characteristics of slender structures
A semi-submersible is usually a column-stabilized unit, which 2.3.1 Mooring systems
consists of a deck structure with large diameter support col-
umns attached to submerged pontoons. The pontoons may be Mooring systems are compliant systems. They provide resist-
ring pontoons, twin pontoons or multi-footing arrangement. ance to environmental loading by deforming and activating
reaction forces. Mooring systems work as spring mechanisms
Semi-submersibles have small waterplane areas, which give where displacement of the floater from a neutral equilibrium
natural periods (in vertical modes) slightly above 20 seconds, position causes a restoring force to react to the applied loading.
usually outside the range of wave periods except for extreme The tension spring effect of mooring lines derives from two
sea states. This implies that a semi-submersible has small ver- mechanisms:
tical motions compared to a monohull floater. However, its
behaviour in extreme weather requires flexible, compliant — hanging catenary effect – from gravity acting vertically on
metallic riser systems or a hybrid arrangement for this concept. the line
A semi-submersible may be equipped with a variety of moor- — line elastic effect – from elastic stretch over the length of
ing systems similar to a FPSO. the line.
Mooring systems with these two mechanisms are called cate- equipped with a separate hydraulic heave compensation sys-
nary moorings and taut moorings, respectively. tem (i.e. tensioner) to account for the floater motions and at the
same time maintain a constant target value for the applied top
2.3.1.1 Catenary moorings tension. Bending moments are mainly induced by horizontal
Catenary moorings are defined by standard catenary formula- floater motions and transverse loading due to current and wave
tions, which relate the following parameters: submerged action. A pronounced peak in the bending moment distribution
weight of the suspended lines, horizontal mooring load, line is normally seen close to the wave zone.
tension and line slope at fairlead. The compliance to allow for Recently, Spar Supported Vertical Risers (SSVR) have been
wave-induced floater motions is ensured by a combination of proposed and designed for Spar platforms. Top tensions
geometrical change and axial elasticity of the lines. The large applied to the SSVRs are provided by tensioners on the Spar.
line geometrical changes make catenary mooring systems sub-
ject to significant dynamic effects due to transverse drag load. An alternative solution for providing top tension to Spar risers
The mooring lines in catenary mooring systems are commonly is by means of buoyancy modules (air cans) attached along the
composed of steel rope and chain segments. Sometimes clump upper part of the riser inside the moonpool. Several supports
weights and buoys are used to achieve the desired line config- may be placed along the riser system to constrain riser trans-
urations. verse motions. Except for the friction forces there are no con-
straints in riser longitudinal motions. This allows the riser
2.3.1.2 Taut moorings system to move vertically relative to the Spar hull. Bending
moments in risers operated from a Classic Spar are mainly due
In a taut mooring system the lines are nearly straight between to the resulting horizontal hull motions as well as hydrody-
the anchor and fairlead. The vertical forces are taken up as namic loading from the entrapped water in the moonpool. Pro-
anchor and vessel reactions directly. The compliance to allow nounced peaks in the bending moment distribution are
for wave-induced floater motions is provided mainly by line normally found at the support locations.
elasticity.
The transverse geometric changes in taut mooring systems are The static and dynamic behaviour of top tensioned risers is
not as large as in catenary systems, thus dynamic effects due to largely governed by the applied top tension. The effective
transverse drag loads are moderate. weight of the riser system defines the lower bound for the
applied top tension to avoid compression in the riser at static
Synthetic ropes have recently been proposed and used as position. Moreover, a significant higher top tension must be
mooring lines in a taut mooring system to provide required applied to account for imperfect tensioner arrangements and
elasticity and low weight. Compared to steel, synthetic ropes allow for redundancy in case of partial loss of top tension.
exhibit more complex stiffness characteristics (e.g. hysteresis), Increased top tension can also be applied to reduce the proba-
which may induce important dynamic effects. bility of collision in riser arrays and limit the mean angles in
bottom of the risers. The applied top tension is commonly
2.3.1.3 Tendons specified in terms of excess over the effective weight of the
TLP tendons bear much similarity to the mooring lines in a taut riser system, and referred to as overpull. The required overpull
mooring system. However, the fundamental difference is that is system dependent with a typical range of 30-60%.
TLP tendons are usually made of large dimension steel tubes Steel pipes have traditionally been applied for floaters in mod-
that are hardly compliant in the axial direction. The TLP sys- est water depths. With attached buoyancy modules, steel risers
tem acts as an inverted pendulum. The station-keeping forces may be applied for deep water floaters. Titanium and compos-
are governed by tendon length and the pretension. Tethers ite risers are suggested for deep water applications in order to
made of composite material are presently being qualified and keep the top tension requirement at an acceptable level.
will extend the use of TLPs into even deeper waters.
The cross-sectional composition depends on the functional
2.3.2 Riser systems applications. Export, import and low pressure drilling risers
Depending on the mechanism of how floater motions are are normally single tubular risers. Multi-tube cross-sections
absorbed by the riser system, the risers can be divided into the are typically found in high-pressure drilling and workover ris-
following three categories: ers as well as production risers.
Taper joints, flex-joints or ball-joints may be applied to reduce
— top tensioned risers bending stresses at the riser termination at seafloor. Flex-joint
— compliant risers or ball-joint may be applied to reduce bending stresses at riser
— hybrid risers. termination at floater. Taper joint may also be applied at the
keel of Spar and other deep draught floaters.
They are described in the following three sections.
2.3.2.2 Compliant riser systems
2.3.2.1 Top tensioned risers
Vertical risers supported by top tension in combination with Compliant riser configurations are designed to absorb floater
boundary conditions that allows for relative riser/floater motions by change of geometry, without the use of heave com-
motions in the vertical direction are referred to as top tensioned pensation systems. The required system flexibility is normally
risers (TTRs). A TTR is normally constrained to follow the obtained by arranging non-bonded flexible pipes in one of the
horizontal floater motions at one or several locations. Ideally, following ‘classical’ compliant riser configurations; steep S,
the applied top tension should maintain a constant target value lazy S, steep wave, lazy wave, pliant wave or free hanging (cat-
regardless of the floater motions. Hence, the effective tension enary).
distribution along the riser is mainly governed by functional Such solutions will for conventional water depths require a
loading due to the applied top tension and the effective weight. pipe with large capacity regarding tensile loading and external/
The relative riser/floater motion in vertical direction is com- internal pressure combined with low bending stiffness and low
monly termed stroke. Applied top tension and stroke capacity critical radius of curvature, e.g. high ‘volume’ stiffness com-
are the essential design parameters governing the mechanical bined with high bending flexibility.
behaviour as well as the application range. For floaters with The desired cross sectional properties are normally obtained
rather small heave motions such as TLPs, Spar platforms, deep by the introduction of a flexible layered pipe where each layer
draught floaters and semi-submersibles, TTRs can be an attrac- has a dedicated function. The number of layers and properties
tive riser solution. of each layer are selected to meet the design requirements and
TTRs operated from semi-submersibles and TLPs are are hence tailor-made for each actual installation. The vast
majority of flexible pipe designs are non-bonded allowing for without significant damage. To operate permanently, FTLs
relative motions between the layers. need to comply with design requirements for risers.
In deep water, it is also possible to arrange metallic pipes in Load effect analyses of FTLs can be challenging. This is par-
compliant riser configurations. Steel Catenary Risers (SCR) ticularly the case for floating FTLs, which are highly compli-
have been installed in the Gulf of Mexico as well as Brazilian ant due to low effective tension. Furthermore, special load
fields (see e.g. Phifer et al 1994). Steel and titanium risers in models are required to describe variable drag and added mass
Lazy Wave configurations have been proposed for semi-sub- of such systems as the pipe moves in and out of the water when
mersibles and TLPs in deep water. A Lazy Wave configuration exposed to loading from waves and floater motions. Simulta-
with increased horizontal extension termed Long Wave is pro- neous excitation from floater motions at both ends is required
posed for the application of metallic risers for deep water for consistent load effect assessment for rather short FTLs. The
FPSO in North sea conditions (Karunakaran et al 1996). In critical areas for excessive bending/curvature will normally be
such applications it may also be considered to apply pre-bent close to the floater attachments.
pipe sections to reduce the dynamic curvature at critical loca-
tions along the riser, i.e. hog and sag bends. Single pipe cross- 2.3.2.5 Umbilicals
sections are typically applied for compliant riser configura- Umbilicals will normally have complex cross-sectional
tions. designs displaying pronounced nonlinear stiffness characteris-
Compliant riser systems will in general experience signifi- tics, e.g. moment/curvature hysteresis. Umbilicals may be
cantly larger static and dynamic excursions when compared to arranged in the classic compliant riser configurations or
top tensioned risers. The floater motion characteristics will in clamped to a compliant or top tensioned riser. The latter solu-
many situations be decisive for the dynamic tension and tion is commonly termed ‘piggy-back’ and will require special
moment variation along the riser, e.g. TLPs, Semi-submersi- modelling considerations in the global load effect analyses,
bles and ships. Environmental load effects will consequently e.g. evaluation of hydrodynamic coefficients and stiffness
also be of greater concern for compliant configurations. Criti- properties for a double symmetric cross-section. Umbilicals
cal locations on compliant risers are typically the wave zone, are otherwise treated similar to compliant riser systems in the
hog-and sag bends, touch down area at seafloor and at the ter- global load effect analysis.
minations to rigid structures.
2.3.3 Slender structure nonlinearities
Termination to rigid structures are an essential design issue for
compliant riser configurations. Possible solutions are carefully Despite the differences in design, function and application
designed bend stiffener, ball joint or flex joint. The primary areas for the slender structures discussed in the previous sec-
design requirement is to limit bending curvature and pipe tions (top tensioned riser, compliant risers, fluid transfer lines
stresses. The secondary design requirement is to minimise and mooring lines/cables), physical behaviour and governing
forces on the supporting structures. parameters for the response characteristics are quite similar.
Such structures are commonly also termed as tensioned struc-
2.3.2.3 Hybrid riser systems tures to reflect that the effective tension is the overall govern-
ing parameter for the global configuration, i.e. geometry, and
There is significant potential for hybrid riser configurations, transverse stiffness. A common overall analysis framework
combining the properties of tensioned and compliant risers in can be applied in load effect analyses of slender structures.
an efficient way. Most proposed designs are based on combin-
ing a self-supported vertical riser column, i.e. tensioned riser, Mooring lines and cable/chain systems are not influenced by
with a flexible riser at upper end for connection to the floater. bending stiffness. The other systems have a physical bending
stiffness that should be considered in the load effect analyses.
The vertical column is normally governed by a bundle of steel
risers. Control umbilicals may also be integrated in the bundle. Understanding the important non-linearities of slender struc-
A buoyancy module at the upper end provides the required ten- tures is critical for system modelling as well as selection of
adequate global analysis approach. Non-linearities will also be
sion in the riser column. The upper end of the vertical column
is connected to the support floater by several flexible risers. decisive for the statistical response characteristics for systems
exposed to irregular loading. An essential issue is how non-lin-
A major advantage of such designs is that the vertical column ear properties of the slender structure and hydrodynamic load-
is a self-supporting structure. The system can be designed to ing mechanisms transform the wave frequency Gaussian
withstand significant dynamic floater motions since flexible excitation, i.e. waves and 1st order floater motions into non-
risers are used for connecting the floater to the riser column. Gaussian system responses. Important non-linearities to be
However, hybrid riser systems tend to be quite complex struc- carefully considered can be summarised as:
tures with special design challenges. Prediction of the column 1) Geometric stiffness, i.e. contribution from effective ten-
response in severe current conditions requires careful evalua- sion to transverse stiffness. Tension variation is hence a
tion of the hydrodynamic coefficients for the riser bundle. non-linear effect for slender structures.
Evaluation of possible VIV response of the individual tubular
in the riser bundle must also be conducted. 2) Hydrodynamic loading. Non-linearities are introduced by
A special design issue for such systems is the control of the the quadratic drag term in the Morison equation expressed
horizontal floater position relative to the upper column end to by the relative structure-fluid velocity and by integration
avoid excessive loading in the flexible risers. Integrity of the of hydrodynamic loading to actual surface elevation.
subsea buoyancy module is another vital design issue. 3) Large rotations in 3D space. This is relevant for systems
with bending stiffness undergoing two-axial bending.
2.3.2.4 Fluid transfer lines
4) Material and component non-linearities.
Floating/submerged pipes used for transportation of fluids
between two floaters are known as Fluid Transfer Lines 5) Contact problems in terms of seafloor contact and hull/
(FTLs). FTLs are normally low-pressure flexible pipes or slender structure contact (varying location of contact point
hoses. However, use of metallic FTLs has also been proposed. and friction forces).
Buoyancy modules may be applied to achieve a desired config- The relative importance of these non-linearities is strongly sys-
uration for floating as well as submerged FTLs. tem and excitation dependent. Non-linearities due to item 1)
Analyses need to be performed to ensure that FTLs can operate and 2) will, at least to some extent, always be present. Item 3)
safely within defined operational conditions and withstand is relevant for systems with bending stiffness undergoing two-
extreme environmental loading in disconnected conditions axial bending due to in-plane and out of plane excitation, while
4) and 5) are more system specific non-linear effects. Material analyses is just as important as the location of the centre of
non-linearities are important for flexible risers and umbilicals, buoyancy. Influence from potential free surface effects (slack
e.g. hysteretic bending moment/ curvature relation due to tanks) needs to be taken into account while determining the
interlayer stick/slip behaviour, and synthetic mooring lines metacentric height.
(axial force/elongation hysteresis). Component non-linearities The additional restoring effects due to the reaction from the
are experienced for several riser system components such as buoyancy cans on the riser guides also need to be taken into
flex-joint, tensioner, bending stiffener etc. account.
It should be noted that external hydrostatic pressure is not con- Stiffness contributions from moorings lines and risers are
sidered to be a non-linear effect as hydrostatic pressures nor- assumed to be taken into account by the direct FE formulation
mally will be handled by the effective tension/ effective weight in the analyses.
concept (Sparks 1984) in computer programs tailor made for
slender structure analysis (e.g. Engseth et al 1988, O’Brien et The mass distribution of the floater may either be entered as a
al 1988). global mass matrix, or from a detailed mass distribution (e.g.
FE model). The input coordinate system varies depending on
software and may be referred to the vertical centre of gravity,
or the water plane. Input of roll and pitch radii of gyration is
3. Floater Load Models very often a source of error in computer programs. Applying
the correct reference axis system is usually the challenge in this
3.1 General context.
Floater motions are commonly split into LF, WF and HF
motion components. The WF and HF motions are mainly gov- 3.3 Wave loads
erned by inviscid fluid effects, while viscous fluid effects are
relatively important for LF motions. Different hydrodynamic 3.3.1 General
effects are important for each floater type, and must be taken The floaters are usually large volume structures and thus iner-
into account in the analysis and design. An overview of these tia-dominated. This implies that radiation/ diffraction analyses
load effects is presented in Table 3-1. Some of the effects can need to be performed with a suitable analysis tool. Some float-
be linearised and included in a frequency domain approach, ers, such as semi-submersibles and truss Spars, may also
while others are highly non-linear and can only be handled in require a Morison load model for the slender members/braces
time-domain. In comparison with frequency domain analysis, in addition to the radiation/diffraction model.
the advantage of a time domain analysis is that it can easily
capture higher order load effects. In addition, a time domain A linear radiation/diffraction analysis will usually be suffi-
analysis can predict the maximum response without making ciently accurate. The term ‘linear’ means that the velocity
assumptions regarding the response distribution. potential is proportional to the wave amplitude, and that the
average wetted area of the floater up to the mean water line is
In this RP only the hydrodynamic loads that have an effect on considered. The analysis gives first order excitation forces,
the global motions of the floater and its slender structures will hydrostatics, potential wave damping, added mass, first order
be considered. This means that wave in deck loads, slamming motions in rigid body degrees of freedom and second order
loads and green water loads will not be dealt with here. mean drift forces/moments. The mean wave drift forces only
dependent on first order quantities, and can therefore be calcu-
Table 3-1 Hydrodynamic effects of importance for each floater lated in a linear analysis.
FPSO Semi DDF TLP Several wave periods and headings need to be selected such
Wave frequency loads X X X X that the motions and forces/moments can be described as cor-
Low frequency loads X X X X rectly as possible. Cancellation, amplification and resonance
Loads in moonpool X X effects must be properly captured. Modelling principles
related to the fineness of the panel mesh must be adhered to,
Mathieu instability X
e.g.:
Hull vortex shedding X
Wave in deck loads X X X — diagonal length in panel model < 1/6 of smallest wave
Slamming loads X X X length analysed
Green water loads X — fine panel mesh to be applied in areas with abrupt changes
in geometry (edges, corners)
High frequency loads X
— finer panel mesh towards water-line in order to calculate
3.2 Hydrostatic loads accurate wave drift excitation forces.
The structure weight and buoyancy force balance is the starting For radiation/diffraction analyses of FPSOs and Spars atten-
point for hydrodynamic analyses. Influence from risers and tion should be paid to the existence of “irregular frequencies”.
mooring pretensions is part of this load balance. These frequencies correspond to short internal waves in the
numerical model and do not have any physical meaning. It is a
Usually this effort is trivial, but important for the success of deficiency of the mathematical model used. At these frequen-
subsequent hydrodynamic analyses. Buoyancy of large cies a standard sink/source technique may give unreliable val-
volume structures is calculated directly from the wetted sur- ues for added mass and damping. Methods exist to identify the
face geometry described by the radiation/diffraction model. In irregular frequencies. Software SESAM:WADAM provides
cases where a dual model, including Morison elements is features for removing irregular frequencies so that reliable
applied, this may also be handled automatically by the compu- results are obtained for the whole frequency range.
ter program as long as the actual location and dimensions of
the Morison elements are implemented. Hydrodynamic interactions between multiple floaters in close
proximity may also be solved using radiation/diffraction soft-
The moonpool needs some special considerations if the moon- ware through the so-called multi-body options. The n floaters
pool area is large and reduces the waterplane area signifi- are solved in an integrated system with motions in n x 6 DOFs.
cantly. In the case of a Spar with air-can supported riser An example of a two-body system is a LNG-FPSO and a side-
system, using a model with closed bottom of the hard tank or by-side positioned LNG carrier during offloading operations
at keel level will result in too high waterplane stiffness. where there may be a strong hydrodynamic interaction
Applying the correct metacentric height (GML, GMT) in the between the two floaters. The interaction phenomena may be
(RAO), gives the response per unit amplitude of excitation, as time domain analysis methods.
a function of the wave frequency, The radiation/diffraction analysis for a floating structure with
a moonpool should be treated with some care. Moonpool
xWA ( ) H ( ) L ( ) 1
(1 ) (1 )
effects are most relevant for turret moored ships and Spar plat-
forms. Depending on the dimensions of the moonpool, the
where L() is the linear structural operator characterizing the heave motion RAO may be strongly influenced. The motion of
equations of motion, the water in the moonpool has a resonance at a wave frequency
corresponding to the eigenfrequency of an oscillating water
L ( ) 2 M A( ) iB ( ) C
column, Tn 2 h / g where h is the height of the water col-
umn and g is the acceleration of gravity. Neglecting viscous
damping of the water motion in the moonpool will result in
M is the structural mass, A the added mass, B the wave damp- unrealistic large motions and free surface elevation in the
ing and C the stiffness, including both hydrostatic and struc- moonpool close to resonance. Discretization of the wetted area
tural stiffness. The equations of rigid body motions are, in of the moonpool must be done with care in order to capture the
general, six coupled equations for three translations (surge, flow details.
sway and heave) and three rotations (roll, pitch and yaw).
The moonpool effect can be treated in two ways. One approach
The frequency domain method is well suited for systems is to consider the water column motion as a generalized mode.
exposed to random wave environments, since the random Another approach is to consider the motion of a massless lid
response spectrum can be computed directly from the transfer floating on the water column and solve a two-body problem. In
function and the wave spectrum in the following way: both cases additional viscous damping should be introduced.
The damping level can be determined from model tests.
S R ( ) xWA S ( )
(1 ) 2
where ai, aj are the individual wave amplitudes and H(2-) is the
quadratic transfer function (QTF) for the difference frequency
load. The QTF is here presented as a complex quantity with
amplitude |H(2+) | and phase a(2+). Re denotes the real part.
Commercial computer tools exist for calculating the difference
frequency QTF. This is a second-order problem requiring dis-
cretization of the free surface in addition to the floater body
surface.
(1)
The QTFs depend on the first order motions xWA .
The QTF also depends on the directions of propagation i of
the wave components. For short-crested sea-states this means
that it may be necessary to solve the complete bi-chromatic and
bi-directional second-order problem.
QTF satisfies this requirement, while the heave QTF does not.
Using Newman’s approximation to calculate slow-drift forces
significantly reduces computation time since a linear analysis
is sufficient. The diagonal elements H(2-)(i,i) can be calcu-
lated from first-order velocity potential alone. Hence there is
no need to calculate the second order velocity potential.
Newman's approximation usually gives satisfactory results for
slow-drift motions in the horizontal plane since the natural
period is much larger than the wave period. For slow-drift
motions in the vertical plane, e.g. the heave/pitch motions of a
DDF, Newman’s approximation may underestimate the slow-
drift forces and in such case the solution of a full QTF matrix
is required.
For some floater concepts such as TLPs, Newman’s approxi-
mation has been commonly accepted and used in calculation of
slow drift forces/moments due to its efficiency in comparison
with the computation of the full matrix of quadratic transfer
functions (QTF). However, for new floater concepts, caution
should be exercised when applying Newman’s approximation.
It is recommended that the full QTF matrix is computed. It is
especially the case for floaters with relatively large and shal-
low pontoons/bases in relation to the columns. LF roll and
pitch will be the key responses to focus on.
onal elements in the full QTF matrix can be approximated by For floaters like TLPs and Spars it is sufficient to consider
the diagonal elements, i.e. wave drift damping for uncoupled translational modes of
motion (surge, sway). But for FPSOs undergoing large slow
H ( 2 ) (i , j )
2
H (i ,i ) H ( 2 ) ( j , j )
1 (2) drift yaw motions as well, the complete 3x3 wave drift damp-
ing matrix for coupled surge, sway and yaw damping is
needed. In the general case the coupled wave drift damping
Another requirement is that the QTF function is smooth in the forces (Fdx, Fdy) and moment Mdz in the horizontal plane is
region close to the diagonal. Figure 3-1 shows that the surge given by
3.3.4.2 Higher order wave loads tures overriding the aerodynamic damping. For correlation
Deepwater TLPs can experience large resonant high frequency with model test results with only wind loading, the aerody-
transient response, called ringing. Ringing exciting waves namic damping should be estimated and taken into account.
have a wavelength considerably longer than a characteristic 3.5 Current loads
cross section of the structure (e.g. diameter of column). There-
fore, long wave approximations can be applied for higher- Calculation of current loads is challenging due to the fact that
order load contribution. A recommended ringing load model is the current depends on local topographic conditions with often
a combination of full three-dimensional first- and second-order strong variability in magnitude and direction with depth. Only
wave diffraction together with a third-order slender body con- measurements can provide sufficient background for determi-
tribution (Faltinsen et al 1995). Hence, the exciting ringing nation of design current speeds and directions. The current
force can be written as may induce vortex induced motions (VIM) of the floater as
well as vortex induced vibrations (VIV) of the slender struc-
q ( t ) q (WA
1) 2)
( t ) q (WA ( t ) q (FNV
3)
(t)
tures and has to be carefully considered.
A steady current gives rise to a steady force in the horizontal
where qWA(1) (2)
(t ) and qWA (t ) are based on the first and second- plane and a yaw moment. For small displacement floaters in
order force transfer functions. deep water or floaters with a large number of slender struc-
tures, the current loading on the slender structures may domi-
General cubic transfer functions are not yet available so the nate the total steady force. It is therefore of importance to apply
(3)
third-order term, q FNV (t ) is an approximation using a slender the correct drag coefficients with due attention to the excitation
body assumption and is limited to circular column geometries as well as the damping contribution. Sensitivity checks with
in the wave zone. The effect of pontoons on the third-order different sets of drag coefficients are therefore recommended.
term is not included. A validation of this ringing load approach Some recommendations on the selection of drag coefficients
was reported by Krokstad et al (1998). Since ringing is a tran- are included in Appendix A.1.
sient phenomenon, the response must be solved in time
The influence of current on the mean wave drift force is dealt
domain. However, a linear structural model can be applied.
with in 3.3.3.
3.6 Vortex-induced loads
3.4 Wind loads Vortex shedding may introduce cross-flow and in-line hull
Wind loading is important for prediction of global motion motions commonly termed vortex-induced-motions (VIM).
response of floaters. Accurate modelling of the wind effects is Cross flow oscillations are considered most critical due to the
therefore essential. For some floating systems the wind loads higher oscillation amplitude compared to the in-line compo-
can be the dominating excitation. nent.
The global wind loads acting on a floating structure consists of Hull VIM is important to determine as it will influence the
two components, a static part resulting in a mean offset and mooring system design as well as the riser design. Both
mean tilt, and a fluctuating component due to wind gusts which extreme loading and fatigue will be influenced. VIM is a
mainly excite the low frequency motions in surge, sway and strongly non-linear phenomenon and it is difficult to predict
yaw. For some floater concepts the roll and pitch motions are accurately. Model testing has usually been the approach to
also influenced. determine the hull VIM responses. More details can be found
Due to its importance, the wind loading is usually determined in Appendix A.1.
based on wind tunnel tests. These tests are very often con- Floaters with single columns like Spars and multicolumn deep
ducted early in the design process. In case of significant draught floaters are most likely to be exposed to VIM oscilla-
changes to the deck/topside structures during detail design, tions. Therefore, these types of floaters are designed with vor-
these wind tunnel tests may have to be repeated. For minor tex shedding suppression devices like strakes. The inclusion of
deck/topside changes, updates of the wind loading may be per- strakes makes it challenging to perform CFD simulations as it
formed by spreadsheets. will require simulation of 3-dimensional effects, and this
Wind tunnel tests usually cover a sufficient number of wind increases the simulation time considerably. One alternative to
directions such that interpolations can be made in subsequent CFD simulations is to use results from a bare cylinder and use
coupled analyses. The influence of heel may have to be taken empirical data to estimate the reduction in oscillation ampli-
into account if the resulting heel angle is critical and the wind tude due to the strakes. Full-scale data is, however, the ultimate
loading increases considerably with heel angle. This is also solution and should be used to correlate with analytical predic-
needed for floating stability calculations. tions.
The gust wind-loading component is simulated by the wind The most important parameters for hull VIM are:
gust spectrum. A number of wind spectra exist. It should be — A/D ratio (A = transverse oscillation amplitude, D = hull
emphasised that a wind spectrum is selected that best repre- diameter)
sents the actual geographical area the floater is located. Wind — Vr – reduced velocity (= Uc/(fnD), Uc = current velocity, fn
spectra are generally described with a number of parameters = eigenfrequency in transverse direction, D = hull diame-
making it relatively easy to make input errors. Checking of ter).
wind spectrum energies and shapes is therefore considered
essential. The most commonly used wind spectra are the API Typically VIM oscillations will be small and in-line with the
and NPD spectra. Details on these gust wind spectra may be current flow for Vr < 3~4. For Vr > 3~4 the hull will start to
found in the relevant literature. The existence of wind squalls oscillate transverse to the current flow and increase in magni-
requires special attention in those areas it is occurring. tude compared to in-line. Another important effect from the
The wind velocity may be a magnitude higher than the floater transverse oscillations is that the mean drag force increases.
velocity. The use of relative velocity formulation compared to This is also confirmed by model tests and full scale measure-
wind velocity alone will therefore have marginal influence. It ments.
is, however, recommended to use the relative velocity formu- The in-line drag coefficient can be expressed as:
lation also for wind loading. In coupled analyses the aerody-
namic damping contribution is usually insignificant. This is Cd = Cdo[1 + k (A/D)]
due to the larger damping contributions from the slender struc- where
Cdo = initial drag coefficient including influence of strakes floater motions (i.e. correlation in time is maintained). The
k = amplitude scaling factor. floater motions produced by coupled/de-coupled analyses will
A/D = cross-flow amplitude/hull diameter contain combined WF and LF components (e.g. FPSO, TLP,
Spar). TLP motions may in addition contain HF components
The amplitude scaling factor is normally around 2. For a while Spar motions may contain hull VIM components. The
reduced velocity around 5, A/D can be up to 0.7-0.8 if the hull latter will however be in the LF regime due to lock-on to surge/
has no suppression devices such as strakes. Strakes effectively sway eigenfrequencies.
reduce the VIM response down to A/D ~ 0.3 - 0.4.
The described approach is applicable to nonlinear as well as
The coupled analysis approach can be an effective way of linearised TD analyses, but can not be applied in FD analyses.
checking out the responses in moorings and risers by introduc-
ing the known (analytical, model tests, or full-scale) in-line 4.1.2 Transfer function representation
and cross-flow oscillations as forces/moments onto the floater. Slender structure analyses have traditionally been performed
considering dynamic excitation from WF floater motions rep-
Since the vortex shedding is more or less a sinusoidal process, resented by floater motion transfer functions (RAOs). LF
it is reasonable to model the cross-flow force imposed on the motions are considered as a quasi-static effect and accounted
hull as harmonic in time at the shedding frequency fs. VIM for by an additional representative offset, i.e. in addition to
lock-in occurs when the vortex shedding frequency locks on to mean floater position for the actual environmental condition.
the eigenfrequency fn . The vortex shedding is dependent on For Spar platforms, this will also involve an additional heel/tilt
the Strouhal number, and is defined by fs = SUc/D, where S is to account for LF motions.
the Strouhal number. The Strouhal number is typically equal to It should however, be noted that the described approach is only
0.2 for a circular cylinder. In general the transverse (lift) force applicable to slender structures that do not respond dynami-
may be written cally to LF floater motions. Combined WF and LF forced
1 floater motions should be considered if the slender structure
qVIM (t ) U c DC L sin( 2f s t ) dynamics is significantly influenced by LF excitation.
2
The RAO representation of the floater motions is applicable in
where CL is the lift force coefficient. The oscillating in-line TD as well as FD analyses.
force is given by the same expression, except that the oscilla-
tion frequency is twice the vortex shedding frequency fI-L = 2fs. 4.2 Fluid kinematics
The in-line VIM response may be in the order of 0.2 times the Fluid kinematics may comprise a significant dynamic loading
cross-flow VIM response. Hence, the hull VIM response on the upper part of deep water riser systems. Direct wave
curves are typically in the shape of a skewed ‘8’ or a crescent loading on mooring lines is however normally of less impor-
(half moon). tance, except if buoys close to the surface are used to obtain the
desired mooring line configuration.
4.2.1 Wave kinematics
4. Slender Body Load Models
Undisturbed wave kinematics is normally based on Airy wave
This section will give an introduction to commonly used load theory. Wheeler stretching may be applied to compute wave
models for analysis of risers and mooring lines of relevance for kinematics in the wave zone. For further details, see e.g.
slender structure analysis in connection with coupled/de-cou- Gudmestad (1993).
pled system analyses. For a more detailed discussion of special
load models for risers (e.g. slug flow, multi-pipe modelling, 4.2.2 Disturbed kinematics
riser component modelling, temperature effects etc.) reference
is made to e.g. API RP 2RD and DNV OS-F201. The presence of the floater gives rise to changes in the fluid
kinematics. This disturbance may be determined by the use of
4.1 Forced floater motions radiation/diffraction analysis. The outputs from such analysis
are RAOs for disturbed kinematics consistent with the floater
Forced floater motions represent a primary dynamic loading motion RAOs. For floaters and risers located close to e.g. col-
on riser and mooring systems. Floater motions are applied as umns/pontoons, this disturbance must be accounted for in
forced boundary displacements at fairleads of mooring lines design.
and at all relevant supports of riser systems, e.g. multiple trans-
verse riser supports for Spar platforms. 4.2.3 Moonpool kinematics
Floater motions may be specified in terms of motion time his- Kinematics of the entrapped water in the moonpool area can in
tories or floater transfer functions depending on the floater principle be treated in the same way as the disturbed wave kin-
motion analysis strategy as discussed in the following. ematics, i.e. in terms of transfer functions for moonpool kine-
matics consistent with the hull motion transfer functions. This
4.1.1 Time series representation approach requires that the entrapped water is included in the
Time series is the most general format for representation of hydrodynamic model used to compute the floater motion char-
floater motions in slender structure analyses. Simultaneous acteristics. Such calculations will, however, require a very
time series for translations and finite rotations at one location careful modelling to achieve a realistic picture in case of com-
on the floater gives a unique representation of the rigid body plicated moonpool geometry and/or multiple risers in the
floater motion at any location on the floater. Special attention moonpool. Special attention should be focused on possible
should however be given to the definition of finite rotations to resonant modes of the entrapped water, see also 3.3.2 .
ensure consistency. A simplified model for the moonpool kinematics can be
Simultaneous wave time series will in addition be required for obtained by assuming that the entrapped water follows the hull
consistent generation of wave kinematics in the slender struc- motions rigidly. This formulation is applicable for FD as well
ture analysis. as TD analysis. The latter approach allows for consistent treat-
Floater motion time series can be obtained from coupled/de- ment of moonpool kinematics due to simultaneous WF and LF
coupled analyses or measurements (model tests or full-scale). floater motions.
A major advantage of the time series format is that it allows for Assuming that the entrapped water rigidly follows the hull
consistent description of different frequency regimes in the motions, the hydrodynamic loading in the normal (to pipe axis)
Bwind 2CwU w
Bij 2 Bij ( ) S ( )d
0 5.2.3.5 Sea floor friction
where Bij is the wave drift damping coefficient and S() is the Soil friction leads to reduced tension fluctuations for the por-
wave spectrum. tion of the mooring table in contact with sea floor, causing an
increase of the line stiffness. Simulations have shown that low
5.2.3.2 Mooring line damping frequency tensions and damping forces are barely influenced
Drag forces on mooring lines strongly contribute to slow-drift by presence of soil friction, but it has some effect on wave fre-
damping. The wave frequency motions of the floater have been quency tensions (Triantafyllou et.al. 1994).
shown to strongly increase the low frequency damping. Com-
pared with the line diameter, the transverse motion amplitude
is quite large. Hence the flow is well separated and the drag
force can be expressed as
1
dF Cd D (u U ) u U ds
2
where u is the wave frequency velocity and U is the low fre-
quency velocity. D is the characteristic diameter of the moor-
ing line. Since U << u, the linearised drag force is
1
dF Cd D u Uds
2
Mooring line damping can be estimated by relating the damp-
ing to the energy dissipated along the line by the drag forces. Figure 5-1
Assuming that the mooring line behaves in a quasi-static way, Comparison of different slow-drift damping components as func-
tion of wave height (Molin 1993).
the damping can be related directly to the RAO’s of the floater
as shown by Huse (1986).
5.2.3.3 Viscous hull damping 5.2.4 High frequency response
The contribution to damping from viscous forces acting on the Springing response is usually solved in the bi-chromatic fre-
floater is often the most difficult to quantify and its part of the quency domain for each sum-frequency pair using the sum-fre-
total damping may differ significantly from one structure to quency force transfer function H(2+) and a similar linear
another. For an FPSO in surge motion linear skin friction dom- structural operator L as for the wave frequency response,
inates the viscous forces while for a TLP or semi-submersible except for the damping,
(2)
xWA (i j ) H ( 2 ) (i , j ) L1 (i j ) h(t)
1
b( ) ia ( ) e
i t
d
2
Added mass can be taken as the linear high frequency asymp- which is the Fourier transform of the added-mass and damp-
totic limit. A key element of a springing response analysis is to ing, a ( ) A( ) A and b( ) B ( ) . A is the asymp-
estimate damping of the high-frequency vertical motions. The totic value of the added mass at infinitely high frequency. The
following damping contributions should be considered high frequency limit of the wave damping is zero.
The equations of motion can be written as:
i) radiation damping due to radiated free surface waves (can
be neglected),
ii) viscous damping due to hull skin friction and eddy making m A x D1 x D2 f ( x )
damping,
iii) structural damping, t
h(t ) x ( )d q(t , x, x )
iv) flex element damping,
0
v) soil damping. The frequency dependent added mass and damping can be
obtained from a three-dimensional panel program.
5.3 Time domain analyses
Solving the integral-differential equation above may be very
5.3.1 Formulations time consuming due to the strict requirements on time steps
necessary to capture the wave frequency motions. A common
The equations of motion for a freely floating or moored struc- approach is to use a multiple scale approach and separate the
ture can be written as: wave-frequency part from the low-frequency part. The wave-
frequency part is usually solved in the frequency domain,
M x B x D 1 x D 2 f( x ) Kx q ( t , x, x ) which requires the motions to be linear responses to waves.
This means that the quadratic damping D2 is set to zero and the
where stiffness K is constant.
(1)
The exciting force is separated in a wave-frequency part qWA
M = mass matrix, m+A(), including added mass. and a low-frequency part q = q WI q WA q CU q ext .
(2-) (2)
because frictional effects depends on the displacement history using a non-linear integration scheme will ensure consistent
and hence they are impossible to include via restoring force treatment of floater/slender structure coupling effects, i.e.
characteristics. these coupling effects will automatically be included in the
solution.
5.3.4 Slender structure/floater coupling effects
The floater load model accounts for the mass, hydrostatic stiff-
Separate assessment of slender structure/floater coupling ness, frequency dependent added mass and damping as well as
effects is required due to the simplified representation of slen- excitation from wind, waves and current on the floater. Note
der structures in de-coupled analyses. that the floater model applied in coupled analyses is in general
The force on the floater from current loading on the slender identical to the floater model applied in de-coupled analyses
structures can be assessed by static analyses using a standard (see Section 5). The differences are the solution strategy and
slender analysis computer tool or by a coupled model of the the slender structure representation. The slender structure anal-
total system. ysis computer program is the most computationally intensive
process and hence governs the efficiency and stability of the
The contribution from current loading on the slender structures coupled analyses.
can be applied directly as an additional force on the floater in
de-coupled analyses. The dynamic loading from wind and waves is modelled as sta-
tionary stochastic processes in a coupled analysis. Simulations
LF slender structure damping can be assessed with basis in of 3-6 hours will be required to obtain extreme response esti-
coupled analyses (see 6.3.1) or from model tests. For an outline mates with sufficient statistical confidence. This is of particu-
of techniques for estimation of an equivalent linear damping, lar importance for response quantities with significant LF
reference is given to Ormberg et al (1998) and Astrup et al components.
(2001). The estimated damping coefficient can be included
directly in the linear damping matrix for the floater in the de- The general modelling capabilities of FE slender analysis
coupled analysis. codes can be applied to establish models of complex offshore
installations. It is in principle straightforward to establish cou-
It should however be emphasised that the LF damping from the pled models of systems with two or more moored floaters, pos-
slender structures for some systems is sensitive to the environ- sibly interconnected by e.g. flow lines. Hydrodynamic
mental excitation (wave height, period, current etc). Estimates interaction between the floaters may be accounted for in the
of LF damping should therefore preferably be based on the hydrodynamic radiation/diffraction analyses to establish wave
same environmental condition as considered in the de-coupled forces on the floaters. See 3.3.1 for further details.
floater motion analysis.
6.2 Coupled system analysis
Floater response as well as detailed mooring line/riser
6. Coupled Response Analyses response can be computed by coupled analysis using a detailed
model of the total system. This approach is usually termed
6.1 General methodology ‘Coupled System Analysis’ reflecting that all relevant system
The floater, risers and mooring system comprise an integrated responses are computed directly by the fully coupled response
dynamic system responding to environmental loading due to model. This approach may be suitable for rather simple sys-
wind, waves and current in a complex way. The floater tems where adequate mooring line and riser response can be
motions may contain the following components: predicted by fairly simple FE models.
Selected modelling may be applied to gain computational effi-
— Mean response due to steady current, mean wave drift and ciency for more complex systems. In this approach, detailed
mean wind load models of identified critical slender structures are included in
— WF response due to 1st order wave excitation the coupled model (e.g. 1-2 detailed riser models for critical
— LF response due to wave drift, wind gusts and viscous drift slots); otherwise a rather coarse slender structure model is
— HF response (TLP) applied to reduce the computational efforts. This model will
— Hull VIM (Classic/Truss Spar, Mini TLP, DDF) hence behave as a coupled system analysis as described above
for the selected critical slender structures.
These response components will consequently also be present
in the slender structure response. Furthermore, the WF, LF and 6.3 Efficient analysis strategies
HF components are generally described as stochastic proc-
esses. Coupled analyses normally demand substantial computational
efforts. More efficient computation schemes are therefore
The purpose of a coupled analysis is to accurately predict the needed for use in practical design analyses.
floater motions as well as the mooring and riser system
response with due regard of floater/slender structure coupling Several strategies can be proposed to achieve computational
effects. Such analyses need to be conducted for numerous sta- efficiency. All strategies have in common that the floater
tionary design conditions to cover extreme conditions, fatigue motion and slender structure analyses are carried out sepa-
load cases, accidental conditions as well as temporary condi- rately, with the exception of selected modelling stated earlier.
tions. Furthermore, analysis of several modifications of the The first step is always a floater motion analysis. Computed
design should be foreseen as a part of the design process. floater motions are then applied as loading in terms of forced
Hence, computational efficiency and numerical stability is a boundary displacements in subsequent slender structure analy-
key issue in practical design analyses of floating offshore sis. Critically loaded risers and/or mooring lines can then be
installations. analysed one by one in the slender structure analyses. This
scheme contributes to computational flexibility as well as a
A coupled dynamic model of a floating installation can in prin- significant reduction in computation time.
ciple be obtained by introduction of the rigid body floater
model in a FE model of the complete mooring and riser system. 6.3.1 Coupled floater motion analyses
Such models can be quite complex, and a ‘master-slave’
approach is an efficient technique for connecting relevant 6.3.1.1 Methodology
mooring lines/tethers/risers to the floater. The availability of Coupled floater motion analysis in combination with subse-
beam- and bar elements in the FE code is essential for efficient quent slender structure analysis is generally recommended to
modelling of the slender structures. achieve a more efficient and flexible analysis scheme. A flow
Solution of this coupled system of equations in time domain chart for this approach is shown in Figure 6-1. Through careful
modelling, this approach is capable of predicting floater quasi-static) on the WF response. Such effects may be impor-
motions and detailed slender structure response with the same tant for some deepwater mooring lines and riser designs.
precision as the complete coupled system analysis.
Traditional assumptions can alternatively be applied consider-
The primary purpose of a coupled floater motion analysis is to ing WF floater motion as dynamic excitation while LF floater
give a good description of floater motions; detailed slender motions are accounted for by an additional offset (branch b).
structure response is secondary. It can therefore be proposed The slender structure is consequently assumed to respond
to apply a rather crude slender structure FE model (e.g. crude quasi-statically to LF floater motions.
mesh, no bending stiffness in risers, etc) in the coupled analy-
sis still catching the main coupling effects (restoring, damping 6.3.1.2 Modelling considerations
and mass). The numerical solution technique as well as floater
force model is however identical to the approach applied in the The principle applied to establish an adequate simplified slen-
der structure model will depend on the actual system layout as
coupled system analysis as described in 6.2.
well as the required output from the analyses. The primary
This approach gives a significant reduction in computation requirement is to give an adequate representation of the cou-
time due to a reduced number of degrees of freedom in the cou- pling effects. However, it is also often desirable to establish
pled analyses. Computation times close to real time have been some key results for the mooring and riser system directly as
experienced for quite complex FPSO and Spar systems (Orm- output from the coupled floater motion analyses. Such infor-
berg et al 1998, Astrup et al 2001) mation can be used to identify critically loaded slender struc-
tures to be analysed in more detail.
Advanced In most situations, it is convenient to include all mooring lines,
vessel model tethers and risers in the FE slender structure model. The FE
Vessel Motion model of each slender structure component is simplified to the
Analysis extent possible using a rather rough mesh and omission of
Simplified
bending stiffness for most parts of the riser system. This will
slender structure
allow for output of key slender structure responses, e.g. moor-
model ing line tensions at fairlead, riser top tensions, tensioner stroke
etc., directly from the coupled floater motion analysis.
LF & WF vessel
motions More detailed riser responses requiring a refined FE model of
the riser system are carried out separately in dedicated riser
analyses to save computation time and increase the analysis
flexibility. Examples are modelling of special components
Select vessel motion such as taper joints as well as refined mesh for adequate calcu-
(b) (a) lation of moment, shear and curvature in critical areas, e.g.
representation touch-down area for SCRs.
Further simplifications are possible if the primary purpose of
the coupled analyses is the prediction of floater motion
Establish WF &LF response. These simplifications may involve the use of equiv-
‘representative’ vessel alent models for groups of mooring lines, tethers and risers.
offset (mean & LF) motions This will give a less transparent overview of the slender struc-
ture response but may be fully acceptable for description of the
floater motion response with due regard of coupling effects.
7. References
Advanced Coupled Vessel [1] Aquirre, J.E. and Boyce, T.T. (1974) “Estimation of wind
vessel model Motion Analysis forces on offshore drilling platforms. Trans. Royal Inst. Nav.
(20~25 LF Arch. (RINA), 116.
Simplified motion cycles) [2] Arahna, J. A. P. (1996): “Second order horizontal steady
slender structure forces and moments on a floating body with small forward
model speed”. J. Fluid Mech. Vol. 313.
LF & WF vessel [3] Astrup O. C, Nestegård A., Ronæss M., and Sødahl N.
motions (2001) ”Coupled Analysis Strategies for Deepwater Spar Plat-
forms” Proc. ISOPE 2001
[4] Engseth, A., Bech, A. and Larsen, C.M. (1988) “Efficient
LF damping Method for Analysis of Flexible Risers”. Proc. BOSS 1988.
estimation [5] Faltinsen, O.M. (1990): “Sea Loads on Ships and Offshore
Structures”, Cambridge University Press.
Advanced [6] Faltinsen, O.M., Newman, J.N., Vinje, T., (1995), Nonlin-
vessel model ear wave loads on a slender vertical cylinder, Journal of Fluid
Vessel Motion Mechanics, Vol 289, pp. 179-198.
Analysis [7] Finne, S., Grue, J. and Nestegård, A. (2000) ”Prediction of
Slender structure
the complete second order wave drift damping force for off-
restoring force
shore structures”. 10th ISOPE Conference. Seattle, WA, USA.
characteristics
LF & WF vessel [8] Gudmestad, O. (1993) “Measured and Predicted Deep
Water Wave Kinematics in Regular and Irregular Seas”.
motions Marine Structures. Vol. 6.
[9] Hughes, G. (1954) “Friction and form resistance in turbu-
Select vessel motion lent flow, and a proposed formulation for use in model and ship
(b) (a) correlation”. Transaction of the Institution of Naval Architects,
representation 96.
[10] Isherwood, R.M. (1973) “Wind resistance on merchant
Establish WF &LF ships”. Trans. Inst. Nav. Arch. (RINA), 115.
‘representative’ vessel [11] Karunakaran, D., Nordsve, N.T. and Olufsen, A. (1996)
offset (mean & LF) motions ”An Efficient Metal Riser Configuration for Ship and Semi
Based Production Systems”. Proc. ISOPE 1996, Los Angeles.
[12] Kim, M-S., Ha, M-K. and Kim, B-W. (2003): ”Relative
motions between LNG-FPSO and side-by-side positioned
Advanced slender LNG carriers in waves”. 13th ISOPE Conference, Honolulu.
Vessel WF structure model of
[13] Kim, S., Sclavounos, P.D. and Nielsen, F.G. (1997)
motion RAO each riser & mooring “Slow-drift responses of moored platforms”. 8th Int. BOSS
Conference, Delft.
[14] Krokstad, J.R., Stansberg, C.T., Nestegård, A., Marthin-
Slender structure Slender structure sen, T (1998): “A new nonslender ringing load approach veri-
analysis analysis
fied against experiments”. Transaction of the ASME, Vol. 120,
Feb. 1998
[15] Molin, B. (1993): “Second-order hydrodynamics applied
to moored structures”. 19th Wegemt School, Nantes, 20 – 24
WF slender WF & LF slender Sept. 1993.
structure responses structure responses
[16] Newman, J.N. (1974): “Second Order, Slowly Varying
Forces in Irregular Waves”. Proc. Int. Symp. Dynamics of
Figure 6-2 Marine Vehicles and Structures in Waves, London.
Analysis strategy 2: Vessel motion and separate slender structure
analysis [17] Newman, J.N. (1994): “Nonlinear scattering of long
waves by a vertical cylinder”, Symposium, Oslo 1994.
[18] O’Brien, P.J., McNamara, J.F. (1989) “Significant Char-
The methodology has also been applied to the analysis of Spar acteristics of Three-dimensional Flexible Riser Analysis”.
platforms (Astrup et al 2001). However, less accurate results Engineering Structures, Vol. 11.
were reported due to the complex Spar platform response.
Refer to Astrup et al (2001) for further details. [19] Ormberg H, Fylling I. J., Larsen K., Sødahl N. (1997):
“Coupled Analysis of Vessel Motions and Mooring and Riser
Combined use of coupled/de-coupled analyses is generally System Dynamics”, Proc. OMAE 1997.
considered less accurate than coupled floater motion analyses.
It is however considered as a useful supplement for coupled [20] Ormberg H., Larsen K. (1997): “Coupled Analysis of
analysis, especially for screening purposes and sensitivity Floater Motion and Mooring Dynamics for a Turret Moored
studies. Tanker” Proc. BOSS 1997.
[21] Ormberg H, Sødahl N, Steinkjer O (1998) “ Efficient
Analysis of Mooring Systems using De-coupled and Coupled
Analysis” OMAE 98.
[22] Phifer, E.H., Kopp, F., Swanson, R.C., Allen, D.W. and
Langner, C.G. (1994) ”Design and Installation of Auger Steel forms. Boston Marine Consulting, 1999.
Catenary Risers”. Paper No. 7620, Proc. OTC, Houston. [25] Triantafyllou, M.S., Yue, D.K.P and Tein, D.Y.S. (1994)
[23] Sparks, C.P. (1984) “The Influence of Tension, Pressure “Damping of Moored Floating Structures”. OTC7489 Off-
and Weight on Pipe and Riser Deformations and Stresses.” shore Technology Conference, Houston 1994.
Journal of Energy Resources Technology, ASME, Vol. 106. [26] Huse, E. “Influence of Mooring Line Damping upon Rig
[24] SWIM – Frequency-domain analysis of offshore plat- Motions” OTC Paper 5204, 1986.
APPENDIX A
SELECTION OF DRAG COEFFICIENTS
sions and current velocities) are available, this will certainly Singularities will be introduced if sinks/sources are distributed
add valuable information and validation. on each side of a thin plate with thickness much smaller than
the characteristic panel size. To avoid such problems, the
Table A-1 Measured model and full scale Cd for wire, chain heave plates can be made artificially thicker. The heave added
and risers mass is not sensitive to the thickness for small thickness values
Re Cd Notes
and the contribution to the sway/surge added mass is small
compared to the contribution from the hard tank. An alterna-
Wire, model scale tive is to model the heave plate as a dipole sheet if the wave dif-
11 – 140 2.0 – 1.0 D = 0.65 – 3 mm, towing fraction computer program [A1-5] has this option.
13 – 120 1.1 – 0.9 Scale: S200 – S55 The heave added mass for a square plate in an infinite fluid is
120 – 14000 0.8 – 1.1 D = 1.1 – 3.8 mm, towing given by the formula M a 0.145 b 3 where b is the hori-
Chain, model scale zontal dimension of the plate. (Ref. [A1-1]).
13 – 110 3.0 – 2.5 D = 1.05 mm, towing The vertical motion of the Spar is very small compared to the
13 – 120 2.5 – 1.8 Scale: S200 – S55 horizontal dimension of the heave plate. This means that the
Risers, model scale KC number for periodic motion
120 – 1100 1.4 – 1.15 Equiv. Risers, S200 – S55 z
KC 2 WA
100 2.0 Single J, Scale S150 b
Wire, full scale (zWA is the heave motion) is very small while the Reynolds
104 – 1.4104 1.1 – 0.95 D = 1.1 – 38 mm, drop tests number Re is in the order of 106. The drag coefficient in this
1.4104–1.1105 1.05 – 0.90 D = 78 mm, towing flow regime is very sensitive to KC and increases strongly with
105 0.83 D = 147 mm, vel. = 1 m/s decreasing KC, but is rather insensitive to Re. A formula for
Chain, full scale the drag coefficient of a long, thin plate strip is given in Ref.
[A1-6],
1.410 3 –104 2.7 – 2.1 D = 30 mm, KC = 163 – 306
104–1.3104 2.7 – 2.2 D = 30 mm, towing
1.3104 – 1.1105 2.5 – 1.7 D = 65 mm, towing Cd 15KC 1 / 2 exp(1.88 / Re0.547 )
1.05105 1.4 D = 140 mm, vel. = 1 m/s 3D effects will reduce the drag coefficient. Heave plate drag
Risers, full scale coefficients have been found to be in the order of 5-10 for
1.1105 1.15 D = 200 mm existing Truss Spar designs. It is recommended that model
tests or CFD calculations are performed to verify drag coeffi-
Table A-2 Typical two dimensional drag coefficients, Cd for cients to be used for the heave plates.
Re = 104 – 107. A.5 References
Type Cd range
[A1-1] DNV Recommended Practice DNV-RP-C205, Envi-
Wire, six strand 1.5 – 1.8 ronmental Conditions and Environmental Loads, April 2007.
Wire, spiral, no sheathing 1.4 – 1.6
[A1-2] DNV Recommended Practice DNV-RP-F105 “Free
Wire, spiral with sheathing 1.0 – 1.2 Spanning Pipelines” (2002)
Chain, stud (relative chain diameter) 2.2 – 2.6
[A1-3] Sumer, B.M. and Fredsøe, J. “Hydrodynamics around
Chain studless (relative chain diameter) 2.0 – 2.4 cylindrical structures”. World Scientific, 1997.
Metallic risers 0.7 – 0.9
[A1-4] Sarpkaya, T. and Isaacson, M. “Mechanics of Offshore
Flexible risers 0.8 – 1.0 Structures”. Van Nostrand Reinhold Company, 1981.
A.4 Inertia and drag coefficients for heave plates [A1-5] WAMIT V6.1. A radiation-diffraction panel program
Special attention must be given to the heave plates for a Truss for wave-body interactions. Wamit Inc.
Spar. Heave plates contribute by increasing the vertical added [A1-6] Shih, C.C. and Buchanon, H.J. (1971) “The Drag on
mass and damping forces. The contribution to added mass can Oscillating Flat Plates in Liquids at Low Reynolds Numbers”.
be calculated by modelling the plates in a sink/source analysis. J. Fluid Mech., Vol. 48.