Content uploaded by Lucas Letournel
Author content
All content in this area was uploaded by Lucas Letournel on Dec 04, 2024
Content may be subject to copyright.
STATISTICAL TIME DOMAIN PERFORMANCE
ASSESSMENT OF BOAT LANDING TRANSFER
OPERATIONS
L. LETOURNEL(1), F. RONG`
ERE(1)
C. CHAUVIGN´
E(1), S. KERKENI(1)
lucas.letournel@dice-engineering.com ; francois.rongere@dice-engineering.com
camille.chauvigne@dice-engineering.com; sofien.kerkeni@dice-engineering.com
(1)D-ICE ENGINEERING, Nantes
R´esum´e
Les op´erations de transfert de personnels et d’´equipements, d’un navire de transport
`a une plateforme fixe ou flottante, dans des conditions environnementales complexes, sont
des sujets d´elicats, de par leur nature risqu´ee. Leur planification n´ecessite l’´evaluation
des capacit´es d’un navire `a assurer un transfert sˆur, pour des conditions environnemen-
tales donn´ees. Traditionnellement, les limites en hauteur de vagues sont utilis´ees pour
d´efinir l’operabilit´e des navires, bien que ce crit`ere simplifie `a l’extrˆeme l’impact des
vagues sur le navire [1]. Le consortium ”Carbon Trust” [2] a mis au point des crit`eres
d’op´erabilit´e, en lien avec les r´eponses en mouvement du navire, bas´es sur des essais
en mer standardis´es, pour aider le milieu industriel `a mieux comprendre les diff´erents
facteurs limitant l’op´erabilit´e des navires. La d´efinition de ces crit`eres s’appuie sur des
quantit´es statistiques : pourcentage de glissements de la d´efense du navire le long du
d´ebarcad`ere par nombre de vagues, et moyenne quadratique de la r´eponse en roulis. ´
Etant
donn´e que la mod´elisation de l’op´eration de transfert fait intervenir des ph´enom`enes non
lin´eaires (adh´erence/glissement), l’´evaluation num´erique des performances navire n´ecessite
l’analyse statistique de simulations temporelles d’op´erations de transfert. La nouvelle
m´ethodologie pr´esent´ee ici est ainsi bas´ee sur : la repr´esentation num´erique du navire et
de la plateforme en des jumeaux digitaux, suivi de la simulations d’op´erations de transfert
dans les conditions environnementales sp´ecifi´ees, et enfin l’analyse statistique des r´esultats
de simulations pour ´evaluer les capacit´es du navire et son op´erabilit´e. Un cas d’application
de cette m´ethodologie est pr´esent´e pour un couple de navire/plateforme, ainsi que les per-
formances associ´ees `a une conditions environnementales donn´ees. Un format particulier
de repr´esentation des r´esultats est notamment propos´e pour faciliter la compr´ehension
des performances et aider `a ´evaluer l’op´erabilit´e du navire.
Summary
1
Boat landing transfer operation is a sensible topic due to its inherent nature : trans-
ferring personnel and equipment from a vessel, generally a CTV (Crew Transfer Vessel),
onto a fixed or floating platform, in adverse environmental conditions. Its planning re-
quires assessing the capability of the vessel/boat landing system to ensure a safe opera-
tion, for given environmental conditions. Wave height limits are traditionally the criteria
used to express vessels’ operability, despite oversimplifying the impact of the waves on
the vessels[1]. The Carbon Trust consortium[2] provided CTV performance acceptability
criteria related to motion responses, based on a standardised sea trial program, to help
the industry better understand the factors that limit CTV operations. The definition of
these criteria are based on statistical quantities : percentage of fender slips per number
of waves, and roll root mean square value. Since the modeling of such operation involves
nonlinear phenomenons (stick/slip), the numerical evaluation of performances requires a
statistical analysis on time domain simulations of the transfer operations. The new me-
thodology presented hereafter is then based on : the numerical representation of the vessel
and platform, as digital twins, followed by the simulation of transfer operations for the
specified environmental conditions, and finally the statistical analysis of the simulation
results to assess the performance of the vessel and its operability. The application of the
methodology is presented for a CTV/platform couple, along with the statistical results
of its performance in specified conditions. A particular plot format is also put forward to
help assess its operability.
2
I – Introduction
The installation and maintenance of facilities in the offshore wind industry rely hea-
vily on crew transfer vessels (CTV), for the transfer of equipment and workforce. The
planification of the transfer operations is then dedicated to find the best window to en-
sure a safe operation, at minimal cost (minimal delay for personnel and vessel). The first
consideration for evaluating the capability of CTV to perform safely was historically limi-
ted to the sea state’s significant wave height (usually between 1 meter, and 1.5 meters),
without any real certainties on the behavior of the vessel in these conditions. In order
to provide a better understanding of the CTVs’ behavior and performance, preconisa-
tions were established by the Carbon Trust [2], along with the main factors limiting the
transit and transfer operations. A standardised sea trial program was thus developed to
assess the CTV’s performance according to new criteria, based on the vessel’s motions.
While computer simulations are mentioned in the Carbon Trust’s report, no methodology
is exposed to explain how the results are obtained. The transfer operation is indeed a
complex problem to model and simulate with multi-physics and multi-bodies involved.
In the offshore wind industry, the operation consists in a push-on/step-across procedure,
meaning that the CTV’s operator pushes the CTV’s fender on the boat landing, with
a designed bollard pull load (usually 80% of the Maximum Continuous Rating MCR),
during a relatively short period (10 minutes). The thrust is supposed to ensure that the
CTV’s bow remains stationary, with no slides along the boat landing, allowing a safe
transfer of personnel onto the ladder. The modeling must include the hydrodynamic loads
and interaction between the CTV and the platform and the contact of the CTV’s fender
on the boat landing, including the stick/slip and elasticity phenomena.
The publications available in the literature concerning the modeling and simulation
of transfer operations focus on specific aspects. Ferreira Gonz´alez et al. [4] investigated
the landing manoeuvre experimentally and numerically by comparing results of time do-
main simulations of a linear potential flow solver (with a simplified fender contact force
model), with wave basins tests at model scale. While the mechanical characteristics of
the fender were obtained experimentally, no boat landing was considered : the fender
was directly in contact with the monopile. Despite tests being carried out in irregular
waves, the evaluation of the CTV’s performances, with respect to criteria, was also not
part of the study. Guanche et al. [5] provided a methodology for the assessment of walk-
to-work accessibility, so a different method to transfer personnel, which does not require
to model the stick/slip phenomenon. The modeling was then based on a rigid, constrai-
ned multibody hydrodynamic model in frequency domain, with linearised mooring and
viscous damping forces. K¨onig et al. [6] focused on a coupling strategy for the different
subproblems to be solved. They proposed a partition solution strategy, based on existing
software platform and solvers. More recently, Otsubo[8] and Meyer[7] studied respectively
the stick/slip phenomenon and the influence of a fixed monopile on the CTV, in terms of
diffraction. While these publications helped us clarify the modeling and simulation tools
for transfer operations, we did not find any detailing a suitable methodology to evaluate
CTVs’ performances according to the Carbon Trust’s criteria.
This paper is thus dedicated, not only to the modeling and simulation tools we used,
but also to the methodology we developed. An anonymized application case is then presen-
ted, illustrating the results obtained with the simulation tools, but also their post-process
for the evaluation of the CTV’s performances, in several environmental conditions.
3
II – Methodology
The main motivation of the development of this methodology was to find a way to
assess a CTV’s performances. Their evaluation usually goes by comparing behavior res-
ponses to thresholds, according to criteria.
II – 1 Criteria definition
The Carbon Trust established 2 main criteria to assess the safety of the transfer
operation :
•a friction or sliding criteria : 95% of waves must pass without any fender’s slid
above 300mm (one ladder rung)
•a roll criteria : the RMS value must not exceed 3°
We introduced an additional criteria to cope for numerical limits of our model (see
section II – 2), consisting in comparing the lateral loads applying to the vessel to the
bollard pull.
We also reversed the friction criteria, in order to consider the number of slips per
waves, rather than the number of waves without slip.
The three criteria can thus be written :
Γf=Nslips>0.3
Nwaves
<5% (1)
Γr=RMS(roll)<3o(2)
Γl=RMS(F Yhydro)
BP % (3)
II – 2 Numerical modeling
The transfer operation, as performed in the offshore wind industry, relies on the push
during transfer method : the vessel pushes with a designed bollard pull, usually 80% of
the MCR, to ensure a minimal occurrences of fender slips on the boat landing. In order
to model precisely this highly non linear stick/slip phenomenon, time domain simulations
must be considered.
FRyDoM simulation framework
The simulations are computed using FRyDoM[3] (Flexible and Rigid Body Dynamic
Modelling for Marine Operations), an open-source multi-body dynamics and multi-physics
simulation framework, dedicated to complex systems modelling and simulation in a marine
environment, in the time domain. Based on Project Chrono[9], an open source multi-
physics simulation engine, it embeds a collision detection system and contact solver to
take into account impact and frictions due to contacts between bodies.
Dedicated to marine operations and ship manoeuvring, FRyDoM embed propulsion
systems, actuators and control systems. Written in full object-oriented C++ 11/14 and
designed from scratch with an open API, it allows to add new features and models ea-
sily. All modules were validated against the literature, with benchmark tests and results
available respectively in the open-source repository and theory documentation [15].
Hydrodynamic loads, including interactions, are modeled through the potential flow
theory, with the diffraction, radiation and excitation loads coming from our in-house fre-
quency potential flow solver, Helios[10]. Second order mean wave drift loads are computed
on the vessel only, also using Helios with the far field method. Ideally, in the frequency
4
flow solver, the vessel must be considered in its static pitch equilibrium, when pushing
against the boat landing (the propulsive thrust and the boat landing reaction must be
balanced by the hydrostatic pitch restoring load) in the absences of waves.
Environment modeling
The marine environment is considered composed of a flat seabed with finite water
depth, and an irregular wave field. The wave field is modelled using linear Airy wave
theory, with a directional JONSWAP wave spectrum. Directional spreading follows a cos2s
directional law. Waves lengths are adapted to the water depth. Frequency and directional
bandwidth are computed automatically, according to respectfully the peak enhancement
factor and the directional spreading, so that the spreading function of the bandwidth
extremas are below 0.01%.
Mechanical modeling
The mechanical interaction between the fender and the wind turbine transition piece
(WT TP) is realised using a contact detection algorithm and a constraint solver within
a non-smooth contact formulation, with neither energy restitution nor contact flexibility.
The contact follows the Coulomb law for friction. Contact boxes are placed to take into
account the presence of the boat landing tubes and the fender, as seen in Figure 1. The
contact boxes follow the bodies in their motions.
Figure 1 – Collision boxes representation, with their envelope
The wind turbine TP is modelled as a fixed rigid body. The CTV is modelled as a
rigid body with 4 degrees of freedom (surge, heave, roll and pitch). Motion equations
are fully nonlinear. Following the need for simulation stabilization, the sway and the yaw
motions are constrained. The propulsion system is modelled with constant forces applied
at the location of the vessel propellers and the total thrust is equally spread among the
propellers. The direction of the thrust is aligned with the body longitudinal axis, with
propellers supposed to be always immersed.
Hydrodynamic load modeling
Different complementary external hydrodynamic loads are applied to the CTV.
•Hydrostatics : Linear hydrostatic stiffness is applied.
•Wave excitation : Both Froude-Krylov and diffraction effects are summed. They are
obtained from hydrodynamic databases generated using linear potential methods
(Helios). They take into account interactions between the CTV hull and the WT
TP.
5
•Radiation forces : The radiation forces are calculated in time-domain using the
Cummins equation. Retardation functions are computed using hydrodynamic da-
tabases for first-order wave interaction. To the same extent as wave excitation
forces, this model takes into account hydrodynamic interactions between the CTV
and the wind turbine pile.
•Mean wave drift : The second order wave drift forces are computed using the far-
field method on the CTV only, in Helios.
•Linear and quadratic extra roll and/or pitch dampings : modelling of viscous and
turbulent flows effects, not modeled by the linear potential theory, and estimated
from experimental or CFD roll and pitch decays.
II – 3 Statistical analysis
The three criteria being expressed as statistical quantities, the evaluation of the per-
formances must involve a statistical analysis of motions responses of the fender’s slip on
the boat landing, relatively to the waves passing, and the CTV’s roll. The CTV’s res-
ponses depend on operational conditions (bollard pull, loading condition, etc.) that can
be considered constant during the simulations, and environmental conditions (waves).
While wind and current are supposed stationary, the waves is modeled as irregular to
replicate sea states. Traditionally sea states are supposed statistically converged after 3
hours of simulation. However, transfer operations are generally limited to short durations
(≈10 min). Thus, we chose to simulate multiple 10 minutes operations, for the same sea
state, but several sets of wave phases.
The friction, lateral and roll criteria are then statistically analysed to obtain their
main characteristics : mean value, standard deviation, first and last deciles, min and max
values. Since the Carbon Thrust does not mention statistical analysis, the statistical mean
value is used to compare to the acceptability value given by the Carbon Thrust for each
criterion. The other statistical characteristics are also shown for a more comprehensive
understanding.
The statistical convergence was proven to be attained for a minimum of 400 simula-
tions, for one sea state, with on waves mean direction, see Figure 2, using the application
case presented in the following section, in its nominal conditions (bollard pull, front waves,
for the specified sea state case).
III – Application
The methodology presented previously is applied to a catamaran CTV and a monopile
wind turbine transition piece, on one sea state case. The setup is presented first, followed
by time domain simulation results, and finally Performance Plots (P-Plots) as statistical
results.
III – 1 Application setup
The dimensions of the catamaran CTV and the monopile are presented in Table 1 and
Figure 3.
Static and dynamic friction coefficients, between the fender and the boat landing were
specified as respectively 0.8 and 0.6, according to the ones measured and reported in [4].
The wave excitation, radiation and wave drift force loads require the computation
of their respective contributions for each wave component. These computations are per-
6
(a) Friction criterion
(b) Lateral criterion
(c) Roll criterion
Figure 2 – Statistical convergence for the different criteria (mean value (left) and stan-
dard deviation (right)).
Figure 3 – CTV and monopile configuration plan
7
characteristics value symbol units
CTV
length Lpp 25 m
breadth B 9 m
draft T 1.5 m
displacement ∆ 100 tons
bollard pull BP 17.5 tons
fender width Bf0.6 m
inertia
Rxx 0.35B tons.m²
Ryy 0.25L tons.m²
Rzz 0.25L tons.m²
hydrostatic
K33 940 kN/m
K44 10 500 kN.m
K55 32 500 kN.m
K35 700 kN
WT TP diameter DP6 m
depth H 14 m
Table 1 – Main characteristics of the CTV and the monopile
Figure 4 – Mesh of the wetted parts of the CTV (left) and the WT pile (right)
formed using the diffraction-radiation code Helios and the contributions are stored in
hydrodynamic databases. Finally, the interaction cannot be considered for the compu-
tation of the waves drift loads, since the far-field method is used in Helios. Thus, two
computations are required :
•One with the CTV only, in order to evaluate the mean wave drift loads, see Figure 4
(left) ;
•Another one with both the CTV and the WT pile for computing the added mass
and damping coefficients along with the excitation loads by taking into account the
hydrodynamic interactions between the two bodies, see Figure 4 (left and right).
A mesh convergence was carried out to ensure the accuracy of the different contri-
butions. The RAOs resulting from the hydrodynamic database for the CTV only are
displayed in Figure 5 in heave and Figure 6 in pitch, for different wave directions.
An additional quadratic roll damping was also considered to cope for the potential
flow theory limits, with µϕϕ = 1.5E6 kg.m².
8
Figure 5 – Amplitude (top) and phase (bottom) of the RAO in heave of the CTV only.
The vertical black dash line represents the natural frequency for this degree of freedom.
Figure 6 – Amplitude (top) and phase (bottom) of the RAO in pitch of the CTV only.
The vertical black dash line represents the natural frequency for this degree of freedom.
9
III – 2 Time domain results
Once all parameters are correctly setup, in FRyDoM simulation framework, time do-
main simulations are ran. The fender vertical position is then extracted, along with the
waves elevation at the fender location.
An example of the time domain variation of both quantities is shown in Figure 7, for
a sea state specified as : HS= 1.75m,TP= 11.3sand W A = 0 degrees (front waves).
The chosen set of wave phases, among the 400 simulations needed to get the statistical
convergence, is particularly critical, since it yields a 11% friction criteria. Despite long
stick periods (between 30 and 300 seconds, and 350 and 550 seconds), slip occurrences
coincide with large wave elevations at the fender position. The fender is even not capable
of sticking on the boat landing for the full last minute of simulation. As for the start
of the simulation, the fender seems to stabilize at a different position than its supposed
equilibrium position, but it might be due also to the large wave elevation.
Figure 7 – Time domain results for the fender vertical position (blue) and wave elevation
(orange) at the fender position, for the specified sea state.
III – 3 Performances, as statistical results
The statistical CTV’s responses are computed on the 400 simulations for the same
sea state (HS= 1.75m,TP= 11.3s), for each wave direction, and the three criteria. The
results are presented in Figure 8, in terms of :
•mean value : between the orange and blue domains,
•standard deviation : represented by the black segments,
•first decile : bottom of the orange domain,
•last decile : top of the blue domain,
•min and max values : represented by the extremities of the red segments.
The acceptability limits are represented by the black dashed lines, except for the roll
figure, which limit exceeds the plot ranges. An arbitrary 15% limit is given for the lateral
criterion, as an indication of the numerical limits inherent to the modeling. The sway and
yaw DOFs are indeed numerically locked to ensure that the CTV remains in contact with
10
Figure 8 – of the friction (left), lateral (middle) and roll (right) criteria for a varying
incident wave angle
the boat landing during the whole simulation. This may result in better performances for
simulations with side waves than in regular head waves.
As can be seen in Figure 8, the friction criterion mean values are always below the
acceptable limit. We can even point out that for 90% of runs, which correspond to the
last decile value, the friction criterion is below 8% for head waves, but almost reaches
10% for following waves. Performances in beam waves, relatively to the friction criterion,
seem better, but that may be due to the artificial numerical constraints in sway and yaw.
As expected, the catamaran CTV has a very good roll response, with mean RMS values
below 1 degree, mostly for beam waves. The limit for the lateral criterion is exceeded in
beam waves, with mean values reaching 20%.
These P-Plots show that the CTV’s operability could be extended to 1.75m Hs sea-
state, for wave directions close to head waves, rather than the classical 1.5m.
IV – Conclusions
A methodology for the evaluation of CTV’s performances in boat landing transfer
operations was presented. It is adapted to the offshore wind industry’s procedures, in
which the operation is performed according to the push-on/step-across method. The per-
formances are evaluated according to the Carbon Trust’s criteria, in friction and roll,
and an additional lateral criteria, introduced for numerical reasons. Since the stick/slip
phenomenon of the fender on the boat landing is capital, yielding a highly non linear
multy-bodies and multi-physics problem to be solved, time-domain simulations are put
forth, using the FRyDoM framework. The performances are then assessed statistically,
on 400 simulations for the same operational and environmental conditions, but different
sets of wave phases. The statistical quantities (mean value, standard deviation, first and
last deciles, min and max values) are then computed according to each criterion definition
and compared to their acceptability limits. A particular polar plot format, called P-Plots,
was proposed to visualize all statistical quantities for different wave directions.
The methodology was applied to a catamaran CTV in contact with a fixed monopile
11
wind turbine transition piece. The setup of both bodies, including hydrodynamic database
and contact model parameters were presented. Performances of the CTV were shown in
the P-Plots format, for one sea state condition.
Several improvements of the modeling can be suggested : the current contact handling
method, based on the non-smooth Project Chrono approach is actually outdated. Fender
deformation could be modeled using smooth contacts method. The propulsion is currently
modeled as a constant thrust, and could be modeled more precisely and controlled along
with other steering actuators to ensure that the CTV remains in contact with the boat
landing. It would then be possible to get rid of the sway and yaw constraints. The lateral
criterion could be changed to one more fitting to the actual operational conditions. All
these computational modules (propulsion models, control strategy, etc.) are actually deve-
loped at D-ICE Engineering, and already in use within the FRyDoM framework (e.g. 3rd
level Dynamic Positioning capability evaluation). Another important phenomenon which
could be interesting to model is the propulsion ventilation, when the propeller exits the
water, and the thrust is no longer constant.
R´ef´erences
[1] C. Anderberg. Challenges of achieving a high accessibility in remote offshore wind
farms how will changing operational requirements affect access strategy during the
operation and maintenance phase ? Master’s thesis, Chalmers University of Techno-
logy, 2015.
[2] Carbon Thrust. Crew transfer vessel (ctv) performance plot (p-plot) development
notice to the offshore wind energy sector. 2017.
[3] D-ICE Engineering. Frydom : Flexible and Rigid body Dynamic modelling for Marine
operations.
[4] D. Ferreira Gonz´alez, M. Lemmerhirt, M. Abdel-Maksoud, M. K¨onig, and A. D¨uster.
Numerical and experimental investigation regarding the landing manoeuvre of a ca-
tamaran vessel at an offshore wind turbine in waves. In Proceedings of the ASME
2015 34th International Conference on Ocean, Offshore and Arctic Engineering, St.
John’s, NL, Canada, volume 7, 2015.
[5] R. Guanche, M. Martini, A. Jurado, and I. J. Losada. Walk-to-work accessibility
assessment for floating offshore wind turbines. Ocean Engineering, 116 :216–225,
2016.
[6] M. K¨onig, D. F. Gonz´alez, M. Abdel-Maksoud, and A. D¨uster. Numerical investiga-
tion of the landing manoeuvre of a crew transfer vessel to an offshore wind turbine.
Ships and offshore structures, 12(sup1) :S115–S133, 2017.
[7] J. Meyer, T. Grotebrune, M. Wynants, A. Hildebrandt, and T. Schlurmann. The
influence of wave diffraction on the motion of a crew transfer vessel behind a monopile.
In International Conference on Offshore Mechanics and Arctic Engineering, volume
87820, page V05AT06A013. American Society of Mechanical Engineers, 2024.
[8] K. Otsubo. Stick/Slip Phenomenon of a Crew Transfer Vessel Pushing Its Bow
Against an Offshore Wind Tower During a Transfer Operation. International Journal
of Offshore and Polar Engineering, 33(04) :409–419, 12 2023.
[9] Project Chrono. Chrono : An Open Source Framework for the Physics-Based Simu-
lation of Dynamic Systems. Accessed : 2016-03-07.
12
[10] P.-Y. Wuillaume, L. Letournel, F. Rong`ere, and C. Chauvign´e. Use of a linear discre-
tization of the velocity potential in the frequency-domain linear potential flow theory.
In International Conference on Offshore Mechanics and Arctic Engineering, volume
85895, page V05AT06A045. American Society of Mechanical Engineers, 2022.
13
