Access to this full-text is provided by MDPI.
Content available from Journal of Marine Science and Engineering (JMSE)
This content is subject to copyright.
Citation: Ivanov, G.; Hsu, I.-J.; Ma,
K.-T. Design Considerations on
Semi-Submersible Columns, Bracings
and Pontoons for Floating Wind. J.
Mar. Sci. Eng. 2023,11, 1663.
https://doi.org/10.3390/
jmse11091663
Academic Editor: José-Santos
López-Gutiérrez
Received: 6 July 2023
Revised: 11 August 2023
Accepted: 21 August 2023
Published: 24 August 2023
Copyright: © 2023 by the authors.
Licensee MDPI, Basel, Switzerland.
This article is an open access article
distributed under the terms and
conditions of the Creative Commons
Attribution (CC BY) license (https://
creativecommons.org/licenses/by/
4.0/).
Journal of
Marine Science
and Engineering
Article
Design Considerations on Semi-Submersible Columns,
Bracings and Pontoons for Floating Wind
Glib Ivanov * , I-Jen Hsu and Kai-Tung Ma
Department of Engineering Science and Ocean Engineering, National Taiwan University, Taipei 106319, Taiwan;
ijenhsu@ntu.edu.tw (I.-J.H.); kaitungma@ntu.edu.tw (K.-T.M.)
*Correspondence: f10525106@ntu.edu.tw
Abstract:
Floating offshore wind turbine (FOWT) is an innovative technology with little industry
guidance for its hull design. Various FOWT floaters with different hull shapes claim to support
the same turbines. Structural integrity and material expense analyses of different pontoon shapes
were conducted, and it was found that some configurations, such as those with every two columns
connected by both pontoon and bracing, have advantages over others. However, it is important to
note that the choice of pontoon shape should be based on the wave loading conditions the floater
will be exposed to. While a T-shaped pontoon provides a cost-effective solution under certain
wave loading scenarios, it may not be the best option for all conditions. Specifically, ring pontoon
designs with full bracing were found to be necessary for withstanding certain wave loads. Therefore,
it is important to consider different Dominant Load Parameters (DLP) and ensure that a FOWT
floater can withstand all applicable DLPs. An uneven hexahedral column shape, which combines
the best attributes of square and round shapes, is proposed as a better alternative to cylindrical
columns. It offers ease of manufacture and reasonably low drag. Bracing is found to be necessary for
withstanding the wind turbine’s incurred moment and forces. The conclusion is that platform design
should prioritize manufacturing costs and strength over maximizing hydrodynamic performance.
Keywords:
floating offshore wind turbines; FOWT; hull design; wave load; bracing; structural
analysis; scalability; manufacturing; offshore wind
1. Introduction
Recently, there have been many different semi-submersible floating offshore wind
turbine designs introduced to public attention [
1
]. As with every new technology, it is
uncertain which option is the best, only with time will natural evolution strike out inefficient
designs and standardize the invention to some extent. Regarding semi-sub FOWT, some
configurations have already been well established in the industry, such as tri-column design
and off-center turbine placement, as seen in commercialized projects. However, other parts
of the semi-sub design are still highly debated, such as how the pontoon should look and
the optimal column shape.
Among current and planned projects by different companies and research teams
(Figure 1), there are V-shaped (Fukushima Shinpuu) [
2
], T-shaped (Bassoe T-floater) [
3
],
above-water pontoons (Gusto Tri-Floater old version) [
4
], tubular joints (WindFloat) [
5
] and
ship-like hull pontoons (DeltaFloat) [
6
], platforms with and without above-water bracing
(EOLFI floater [
7
], Fukushima Shinpuu), platforms with bracing bigger than pontoon
(SanXia YingLing Hao) [
8
], etc. It is evident that every design team drafts their design with
particular constraints and necessities in mind; however, as most are made by for-profit
companies, not much of their thought process is shared with the general public.
As researchers, we take the liberty to analyze and compare them without bias. With
so many varieties, it is clear that some are to be outlived by the more efficient. The parts
specifically looked at in this paper are the different pontoons, columns and, to a lesser
extent, bracing configurations.
J. Mar. Sci. Eng. 2023,11, 1663. https://doi.org/10.3390/jmse11091663 https://www.mdpi.com/journal/jmse
J. Mar. Sci. Eng. 2023,11, 1663 2 of 21
J.Mar.Sci.Eng.2023,11,xFORPEERREVIEW2of22
Figure1.Differentdesignsofsemi-submersiblehullsforFOWT.
Asresearchers,wetakethelibertytoanalyzeandcomparethemwithoutbias.With
somanyvarieties,itisclearthatsomearetobeoutlivedbythemoreefficient.Theparts
specificallylookedatinthispaperarethedifferentpontoons,columnsand,toalesser
extent,bracingconfigurations.
Pontoonandbracinghaveapracticallysimilarpurpose—toholdthecolumnsto-
getherforthemtomakeupaplatform.Indesignssuchaswindfloats,theyarepractically
thesame,theonlydifferencebeingthepontoonislocatedunderwater,whilethebracing
isabovewater.Bracingalsoprovidesawayforworkerstowalkbetweenthecolumns.
Thepontoonisfilledwithballastandthusbringstheplatform’scenterofgravity
down;ifitsverticalfootprintisbig,italsoactsasaheave-dampingplate.Sometimes,it
canalsoprovidebuoyancyaswillbedescribedlater.
2.Methodology
Tocomparedifferentpontoons,10modificationsofabaseplatformAwerecreated
inANSYSSpaceClaim,shownastheshapesBtoKinFigure2,andacylindricalcolumns
shapeL.TheoriginaltriangularringpontoonplatformisanupdatedversionofTaid a Float
[9],aprototypeplatformtobeusedinTaiwa n strait.WhileTaidaFloatpontoonscanbe
fullyfunctional(providebuoyancy)orpartiallyfunctional(permanentlyballasted),they
willbeconsideredonlypartiallyfunctionaltoremovetheinfluenceofwildlyvarying
buoyancybetweenthedifferentpontoons.Theotherdimensionsincludingallcolumndi-
mensionsandoveralllengthandwidtharethesameamongdesigns,asshowninTa ble 1.
Theplatformsupportsa15MWIEAturbine[10,11],withthefollowingcharacteristics
showninTab l e2.Theplatformalsotakessomeballastincolumnsforuseinadynamic
stabilizationsystemandtoachievethedesireddraught.
Figure 1. Different designs of semi-submersible hulls for FOWT.
Pontoon and bracing have a practically similar purpose—to hold the columns together
for them to make up a platform. In designs such as wind floats, they are practically the
same, the only difference being the pontoon is located underwater, while the bracing is
above water. Bracing also provides a way for workers to walk between the columns.
The pontoon is filled with ballast and thus brings the platform’s center of gravity
down; if its vertical footprint is big, it also acts as a heave-damping plate. Sometimes, it can
also provide buoyancy as will be described later.
2. Methodology
To compare different pontoons, 10 modifications of a base platform A were created in
ANSYS SpaceClaim, shown as the shapes B to K in Figure 2, and a cylindrical columns shape
L. The original triangular ring pontoon platform is an updated version of TaidaFloat [
9
],
a prototype platform to be used in Taiwan strait. While TaidaFloat pontoons can be fully
functional (provide buoyancy) or partially functional (permanently ballasted), they will be
considered only partially functional to remove the influence of wildly varying buoyancy
between the different pontoons. The other dimensions including all column dimensions
and overall length and width are the same among designs, as shown in Table 1. The
platform supports a 15 MW IEA turbine [
10
,
11
], with the following characteristics shown in
Table 2. The platform also takes some ballast in columns for use in a dynamic stabilization
system and to achieve the desired draught.
Table 1. TaidaFloat’s principal characteristics.
Length 81.6 m
Breadth 94.2 m
Height 35 m
Draught 22 m
Total displacement 20,300 t
Hull weight 4002 t
CG height 17.35 m from BP
Coordinates Origin Pontoon centroid at base plane (BP)
For every configuration, turbine, ballast and platform assembly weight characteristics
including mass moments of inertia were calculated using ANSYS Mechanical, and they are
shown in Table 3.
J. Mar. Sci. Eng. 2023,11, 1663 3 of 21
Table 2. Wind turbine characteristics.
Name IEA 15 MW
Roter orientation,
Configuration
Upwind,
3 Blades
Rotor Diameter 240 m
Hub height 150 m
Blade mass 65.25 t
Tower mass 1263 t
RNA mass 991 t
J.Mar.Sci.Eng.2023,11,xFORPEERREVIEW3of22
Figure2.Differentpontoonandbracingarrangements.
Tab l e 1.TaidaFloat’sprincipalcharacteristics.
Length81.6m
Breadth94.2m
Height35m
Draught22m
Totaldisplacement20300t
Hullweight4002t
CGheight17.35mfromBP
CoordinatesOriginPontooncentroidatbaseplane(BP)
Figure 2. Different pontoon and bracing arrangements.
J. Mar. Sci. Eng. 2023,11, 1663 4 of 21
Table 3. Design case mass characteristics.
Case A B C D E F G H
System mass (t) 19,450 19,450 20,920 16,410 16,789 16,750 16,750 16,612
Center of mass, CM (m) (X, Y, Z) (0, 6.23, 18.80) (0, 6.27, 18.27) (0, 3.74, 19.36) (0, 8.51, 22.06) (0, 6.62, 21.91) (0, 10.00, 21.17) (0, 10.02, 21.30) (0, 0.82, 24.62)
Principal
inertias about
CM (kg m2)
Ixx 0.5335 ×1011 0.5330 ×1011 0.5301 ×1011 0.4868 ×1011 0.5218 ×1011 0.4696 ×1011 0.4702 ×1011 0.4950 ×1011
Iyy 0.4931 ×1011 0.4923 ×1011 0.5037 ×1011 0.4438 ×1011 0.4460 ×1011 0.4647 ×1011 0.4632 ×1011 0.4984 ×1011
Izz 0.2833 ×1011 0.2816 ×1011 0.3015 ×1011 0.2306 ×1011 0.2456 ×1011 0.2541 ×1011 0.2529 ×1011 0.3178 ×1011
Ixy −0.6734 ×106−0.6725 ×106−0.1504 ×1070.2873 ×1070.1138 ×1070.1171 ×1080.1172 ×1080.3347 ×107
Ixz −0.3047 ×107−0.3057 ×107−0.3480 ×1070.1343 ×108−0.1606 ×1070.4037 ×1070.4086 ×1070.1330 ×108
Iyz −0.9954 ×1010 −0.1003 ×1011 −0.1040 ×1011 −0.9048 ×1010 −0.1004 ×1011 −0.8096 ×1010 −0.8003 ×1010 −0.1068 ×1011
Every design case should be analyzed from the following viewpoints:
•Strength,
•Stability,
•Costs, and
•Practicality.
This paper focuses on strength and cost only. Stability and practicality may be tackled
in future research.
Nowadays, the common approach is to analyze the structural strength of a large
number of designs with FEM. The ANSYS Mechanical Static Structural module was used
for this purpose here. Hull weight and mass moments of inertia were calculated with
a framed, ship-like, inner structure. The inner structure is based on the original inner
structure for configuration A, slightly modified to fit the shape in other cases. It is shown
in Figure 3.
J.Mar.Sci.Eng.2023,11,xFORPEERREVIEW5of22
Figure3.TaidaFloatdimensions,structuralarrangementandaxes.
Asthefirststep,alltheconfigurationsweresubjectedtostaticloads—gravity,water
pressure(fullyballastedpontooncondition),turbineassemblyweightanda0.1Gaccel-
erationinthexandydirectionstoapproximatelyaccountforthewaveloads.
Itisunsurprisingthateveryconfigurationsuccessfullywithstandsthesestaticloads,
withnoimportantdifferencebetweendifferentdesigns.Anexampleoftheresultisshown
inFigure4.
Figure4.TaidaFloatstaticconditionFEManalysis.
Figure 3. TaidaFloat dimensions, structural arrangement and axes.
As the first step, all the configurations were subjected to static loads—gravity, wa-
ter pressure (fully ballasted pontoon condition), turbine assembly weight and a 0.1 G
acceleration in the x and y directions to approximately account for the wave loads.
J. Mar. Sci. Eng. 2023,11, 1663 5 of 21
It is unsurprising that every configuration successfully withstands these static loads,
with no important difference between different designs. An example of the result is shown
in Figure 4.
J.Mar.Sci.Eng.2023,11,xFORPEERREVIEW5of22
Figure3.TaidaFloatdimensions,structuralarrangementandaxes.
Asthefirststep,alltheconfigurationsweresubjectedtostaticloads—gravity,water
pressure(fullyballastedpontooncondition),turbineassemblyweightanda0.1Gaccel-
erationinthexandydirectionstoapproximatelyaccountforthewaveloads.
Itisunsurprisingthateveryconfigurationsuccessfullywithstandsthesestaticloads,
withnoimportantdifferencebetweendifferentdesigns.Anexampleoftheresultisshown
inFigure4.
Figure4.TaidaFloatstaticconditionFEManalysis.
Figure 4. TaidaFloat static condition FEM analysis.
The actual difference among the 12 pontoon designs can only be seen during its opera-
tion in the rough sea, with dynamic loads. According to ABS Rules [
12
], and other major
class societies’ rules [
13
–
15
], offshore structures’ dynamic performance is to be analyzed
with different DLP (Dominant Load Parameters) and DLC (Design Load Cases). They rep-
resent a comprehensive array of parameters such as wind, loading, waves applied together
in the most unfavorable combinations (DLC), and their effect on particular structure parts
and modes of deformation (DLP).
A common practice for a FOWT strength analysis is similar to oil and gas platforms
with the added complexity of turbine–wind interaction:
1.
Environmental conditions of the planned installation site are determined. This usually
involves placing a buoy on-site at least one-two years in advance. A longer time is
preferred, as 50 year return conditions will have to be extrapolated for the extreme
weather survival load cases. As FOWT are unmanned, a less-conservative 50 year
return period is used as opposed to 100 years for oil and gas.
2.
After an initial design geometry and mass distribution are ready, the platform is
subject to hydrodynamic response analysis, which provides Response Amplitude
Operators (RAO) for 6 degrees of freedom. The motion and acceleration RAO peak
values provide guidance on which wave conditions could impart maximal stress to
the platform; the most dangerous wave condition that is expected to occur is referred
to as the design wave. Then, a pressure distribution on the hull below the surface
can be acquired and mapped onto a more detailed structural model for subsequent
FEA. The RAO analysis and load mapping are usually performed with software, e.g.,
Ansys AQWA used in this paper.
3.
Different from oil and gas, wind turbines impart a huge moment load onto the top
part of the floater. When the mooring system and a turbine have been designed, wind
turbine and mooring line loads can be found using coupled wind–wave simulation
software, such as OrcaFlex used in this paper, and applied as loads in FEA software.
4.
The hydrodynamic loads, wind and mooring loads, equipment weight and wave-
induced acceleration are applied to the global FE model, including hull modelled as
shell elements and scantling modelled as beam elements. For a detailed investigation
J. Mar. Sci. Eng. 2023,11, 1663 6 of 21
of areas of interest and fatigue analysis, a set of local models are made with hull and
girders represented by shell elements.
There is different software for FEA and techniques to constrain the model and map the
loads. The authors map loads from AQWA directly to Ansys Mechanical and use springs
at the place of moorings with constraints at the far end. Finally, a stress distribution can be
obtained for assessment with the maximum stress design criteria, as shown in Figure 5.
J.Mar.Sci.Eng.2023,11,xFORPEERREVIEW6of22
Theactualdifferenceamongthe12pontoondesignscanonlybeseenduringitsop-
erationintheroughsea,withdynamicloads.AccordingtoABSRules[12],andotherma-
jorclasssocieties’rules[13–15],offshorestructures’dynamicperformanceistobeana-
lyzedwithdifferentDLP(DominantLoadParameters)andDLC(DesignLoadCases).
Theyrepresentacomprehensivearrayofparameterssuchaswind,loading,wavesap-
pliedtogetherinthemostunfavorablecombinations(DLC),andtheireffectonparticular
structurepartsandmodesofdeformation(DLP).
AcommonpracticeforaFOWTstrengthanalysisissimilartooilandgasplatforms
withtheaddedcomplexityofturbine–windinteraction:
1. Environmentalconditionsoftheplannedinstallationsitearedetermined.Thisusu-
allyinvolvesplacingabuoyon-siteatleastone-twoyearsinadvance.Alongertime
ispreferred,as50yearreturnconditionswillhavetobeextrapolatedfortheextreme
weathersurvivalloadcases.AsFOWTareunmanned,aless-conservative50year
returnperiodisusedasopposedto100yearsforoilandgas.
2. Afteraninitialdesigngeometryandmassdistributionareready,theplatformissub-
jecttohydrodynamicresponseanalysis,whichprovidesResponseAmplitudeOper-
ators(RAO)for6degreesoffreedom.ThemotionandaccelerationRAOpeakvalues
provideguidanceonwhichwaveconditionscouldimpartmaximalstresstotheplat-
form;themostdangerouswaveconditionthatisexpectedtooccurisreferredtoas
thedesignwave.Then,apressuredistributiononthehullbelowthesurfacecanbe
acquiredandmappedontoamoredetailedstructuralmodelforsubsequentFEA.
TheRAOanalysisandloadmappingareusuallyperformedwithsoftware,e.g.,An-
sysAQWAusedinthispaper.
3. Differentfromoilandgas,windturbinesimpartahugemomentloadontothetop
partofthefloater.Whenthemooringsystemandaturbinehavebeendesigned,wind
turbineandmooringlineloadscanbefoundusingcoupledwind–wavesimulation
software,suchasOrcaFlexusedinthispaper,andappliedasloadsinFEAsoftware.
4. Thehydrodynamicloads,windandmooringloads,equipmentweightandwave-
inducedaccelerationareappliedtotheglobalFEmodel,includinghullmodelledas
shellelementsandscantlingmodelledasbeamelements.Foradetailedinvestigation
ofareasofinterestandfatigueanalysis,asetoflocalmodelsaremadewithhulland
girdersrepresentedbyshellelements.
ThereisdifferentsoftwareforFEAandtechniquestoconstrainthemodelandmap
theloads.TheauthorsmaploadsfromAQWAdirectlytoAnsysMechanicalanduse
springsattheplaceofmooringswithconstraintsatthefarend.Finally,astressdistribu-
tioncanbeobtainedforassessmentwiththemaximumstressdesigncriteria,asshownin
Figure5.
Figure5.TaidaFloatAQWAmodelanddynamicpressuredistributionexample.
Figure 5. TaidaFloat AQWA model and dynamic pressure distribution example.
However, this process requires a complete design for internal stiffeners, wind turbine
model, ballast distribution, and most importantly, a complete mooring system and many
iterations loops to obtain the result; therefore, it is rare to see different floaters directly
compared to each other. As such, a simplified approach is used here to compare the
12 floater
shapes and find out their advantages rather than obtain a precise maximum
stress value.
For a structure such as a FOWT floater, the matrix of DLP and DLC combinations
reaches more than one hundred cases. For 12 pontoon configurations analyzed in this paper
that would be more than a thousand cases to investigate. However, not all cases are equally
important; here, we use engineering judgement and compare 12 shapes’ performances at
critical DLPs and the roughest wave load DLC. This method was employed with great
success to reduce the number of analyses when investigating a single platform under
different loads [
16
]. This study is similar in that it picks out separate DLPs to look at
structural responses, but the cases are simplified as many different platforms are analyzed
and compared with each other. The flowchart for the process is shown in Figure 6The
mesh size employed was 600 mm, which corresponds to frame spacing. Case A full FEA
result shown in Figure 7.
As we are looking at stress induced only by a particular DLP, we need to isolate
the load from others when we perform the analysis. The DLP concept is not sensitive to
load origin (e.g., wave load, turbine load or impact load), but only to the force direction;
therefore, we apply isolated forces or pressures for each DLP, and the magnitudes are
chosen to represent pressure caused by a wave of 35 m height, just reaching the platform
top deck.
The most critical is arguably DLP 1; when the wave front is along the platform’s X
axis, then one side column is on the wave top, another side column in the crest, and the
main column is at the middle height, as shown in Figure 8. Difference in buoyancy forces
between the two columns produces a twisting moment on the pontoons linked with the
main column, potentially breaking them. This condition is simulated by the application
of two forces in opposing directions on the top decks of side columns. The boundary
condition is a fixed support at the main column top—farthest from the area of interest,
which is the plane defined by the two side columns.
J. Mar. Sci. Eng. 2023,11, 1663 7 of 21
J.Mar.Sci.Eng.2023,11,xFORPEERREVIEW7of22
However,thisprocessrequiresacompletedesignforinternalstiffeners,windturbine
model,ballastdistribution,andmostimportantly,acompletemooringsystemandmany
iterationsloopstoobtaintheresult;therefore,itisraretoseedifferentfloatersdirectlycom-
paredtoeachother.Assuch,asimplifiedapproachisusedheretocomparethe12floater
shapesandfindouttheiradvantagesratherthanobtainaprecisemaximumstressvalue.
ForastructuresuchasaFOWTfloater,thematrixofDLPandDLCcombinations
reachesmorethanonehundredcases.For12pontoonconfigurationsanalyzedinthispa-
perthatwouldbemorethanathousandcasestoinvestigate.However,notallcasesare
equallyimportant;here,weuseengineeringjudgementandcompare12shapes’perfor-
mancesatcriticalDLPsandtheroughestwaveloadDLC.Thismethodwasemployed
withgreatsuccesstoreducethenumberofanalyseswheninvestigatingasingleplatform
underdifferentloads[16].ThisstudyissimilarinthatitpicksoutseparateDLPstolook
atstructuralresponses,butthecasesaresimplifiedasmanydifferentplatformsareana-
lyzedandcomparedwitheachother.TheflowchartfortheprocessisshowninFigure6
Themeshsizeemployedwas600mm,whichcorrespondstoframespacing.CaseAfull
FEAresultshowninFigure7.
Figure6.Experimentflowchart.
AswearelookingatstressinducedonlybyaparticularDLP,weneedtoisolatethe
loadfromotherswhenweperformtheanalysis.TheDLPconceptisnotsensitivetoload
origin(e.g.,waveload,turbineloadorimpactload),butonlytotheforcedirection;there-
fore,weapplyisolatedforcesorpressuresforeachDLP,andthemagnitudesarechosen
torepresentpressurecausedbyawaveof35mheight,justreachingtheplatformtopdeck.
ThemostcriticalisarguablyDLP1;whenthewavefrontisalongtheplatform’sX
axis,thenonesidecolumnisonthewavetop,anothersidecolumninthecrest,andthe
maincolumnisatthemiddleheight,asshowninFigure8.Differenceinbuoyancyforces
betweenthetwocolumnsproducesatwistingmomentonthepontoonslinkedwiththe
maincolumn,potentiallybreakingthem.Thisconditionissimulatedbytheapplicationof
twoforcesinopposingdirectionsonthetopdecksofsidecolumns.Theboundarycondi-
tionisafixedsupportatthemaincolumntop—farthestfromtheareaofinterest,whichis
theplanedefinedbythetwosidecolumns.
Figure 6. Experiment flowchart.
J.Mar.Sci.Eng.2023,11,xFORPEERREVIEW8of22
Figure7.TaidaFloatglobalFEAmaximumstressdistributionexamples.
DLP2,showninFigure9,iswhenthewavefrontapproachesalongtheYaxis,the
maincolumnonthewavetopandbothsidecolumnsinthecrest,orviceversa.Thiscondi-
tionissimulatedbyapplyingtwoupwardforcesatsidecolumnsandadownwardforceof
doublemagnitudeatthemaincolumn;asimplesupportconditionisappliedtoavertical
lineonthefrontsideofthemaincolumn,asitisarigid,non-deformedregioninthiscase.
DLP3isbendingofthemaincolumn,possiblyinducedbywindturbineload,as
showninFigure10.Thisconditionissimulatedbyapplyingamomenttothecircular
turbinefoundationonthetopofthemaincolumn.ThemomentistakenfromanOrcaFlex
typhoonExtremeSeaStatesimulationdescribedin[17];theboundaryconditionisthree
simplesupportsatcolumnbooms(fairleadposition),farfromtheareaofinterest.
Figure8.Wave-i n ducedtwist—DLP1.
Figure9.Wave-i n ducedshearDLP2.
Figure 7. TaidaFloat global FEA maximum stress distribution examples.
J.Mar.Sci.Eng.2023,11,xFORPEERREVIEW8of22
Figure7.TaidaFloatglobalFEAmaximumstressdistributionexamples.
DLP2,showninFigure9,iswhenthewavefrontapproachesalongtheYaxis,the
maincolumnonthewavetopandbothsidecolumnsinthecrest,orviceversa.Thiscondi-
tionissimulatedbyapplyingtwoupwardforcesatsidecolumnsandadownwardforceof
doublemagnitudeatthemaincolumn;asimplesupportconditionisappliedtoavertical
lineonthefrontsideofthemaincolumn,asitisarigid,non-deformedregioninthiscase.
DLP3isbendingofthemaincolumn,possiblyinducedbywindturbineload,as
showninFigure10.Thisconditionissimulatedbyapplyingamomenttothecircular
turbinefoundationonthetopofthemaincolumn.ThemomentistakenfromanOrcaFlex
typhoonExtremeSeaStatesimulationdescribedin[17];theboundaryconditionisthree
simplesupportsatcolumnbooms(fairleadposition),farfromtheareaofinterest.
Figure8.Wave- i ndu c edtwist—DLP1.
Figure9.Wave- i ndu c edshearDLP2.
Figure 8. Wave-induced twist—DLP 1.
DLP 2, shown in Figure 9, is when the wave front approaches along the Yaxis, the
main column on the wave top and both side columns in the crest, or vice versa. This
condition is simulated by applying two upward forces at side columns and a downward
force of double magnitude at the main column; a simple support condition is applied to a
vertical line on the front side of the main column, as it is a rigid, non-deformed region in
this case.
J. Mar. Sci. Eng. 2023,11, 1663 8 of 21
J.Mar.Sci.Eng.2023,11,xFORPEERREVIEW8of22
Figure7.TaidaFloatglobalFEAmaximumstressdistributionexamples.
DLP2,showninFigure9,iswhenthewavefrontapproachesalongtheYaxis,the
maincolumnonthewavetopandbothsidecolumnsinthecrest,orviceversa.Thiscondi-
tionissimulatedbyapplyingtwoupwardforcesatsidecolumnsandadownwardforceof
doublemagnitudeatthemaincolumn;asimplesupportconditionisappliedtoavertical
lineonthefrontsideofthemaincolumn,asitisarigid,non-deformedregioninthiscase.
DLP3isbendingofthemaincolumn,possiblyinducedbywindturbineload,as
showninFigure10.Thisconditionissimulatedbyapplyingamomenttothecircular
turbinefoundationonthetopofthemaincolumn.ThemomentistakenfromanOrcaFlex
typhoonExtremeSeaStatesimulationdescribedin[17];theboundaryconditionisthree
simplesupportsatcolumnbooms(fairleadposition),farfromtheareaofinterest.
Figure8.Wave- i ndu c edtwist—DLP1.
Figure9.Wave- i ndu c edshearDLP2.
Figure 9. Wave-induced shear DLP 2.
DLP 3 is bending of the main column, possibly induced by wind turbine load, as
shown in Figure 10. This condition is simulated by applying a moment to the circular
turbine foundation on the top of the main column. The moment is taken from an OrcaFlex
typhoon Extreme Sea State simulation described in [
17
]; the boundary condition is three
simple supports at column bottoms (fairlead position), far from the area of interest.
J.Mar.Sci.Eng.2023,11,xFORPEERREVIEW9of22
Figure10.MaincolumnbendingDLP3.
3.PontoonDesign
3.1.FloatersComparison
Forthispartofthisstudy,simplifiedmodelswereused,whichhadauniformplate
thicknessof30mmandnoreinforcement,exceptforbulkheads.Theforceapplication
areaswerekeptrigidusingcoupledmulti-pointconstraints.Theadequacyofthissimpli-
ficationwasdemonstratedinapreviousstudy[9].Bycomparingthestressdistributions
ofafull-detailmodelinFigure7andthesubsequentresults,itcanbeseenthatthestress
concentrationareasarethesame(jointsbetweencolumns,pontoonsandbracings;brac-
ingsplating,thelowestbeltofcolumnplating),thusthesimplifiedmethodandmodels
producecredibleresultsforthiscomparisonstudy.
TheFEAstressdistributionscanbeseeninFigures11–15.Therearequiteafewob-
servations:
(A)Ringpontoonwithbracing (B)Triangularringpontoon
(C)TriangularRing45deg.pontoonwithbracing(D)Y-shapedpontoonwithbracing
Figure11.Cases(A–D)FEADLP1,stressplot.
Figure 10. Main column bending DLP 3.
3. Pontoon Design
3.1. Floaters Comparison
For this part of this study, simplified models were used, which had a uniform plate
thickness of 30 mm and no reinforcement, except for bulkheads. The force application areas
were kept rigid using coupled multi-point constraints. The adequacy of this simplification
was demonstrated in a previous study [
9
]. By comparing the stress distributions of a
full-detail model in Figure 7and the subsequent results, it can be seen that the stress
concentration areas are the same (joints between columns, pontoons and bracings; bracings
plating, the lowest belt of column plating), thus the simplified method and models produce
credible results for this comparison study.
The FEA stress distributions can be seen in Figures 11–15. There are quite a few observations:
DLP 1
: Faced with a beam wave-induced twist, floaters with a pontoon connecting
every two columns resist the load well, while floaters without all three pontoons or with
an intersection in the middle show large areas of stress concentrations (cases F, G, H, and
K). Because the structure is not rigid enough, a small force is able to give birth to a large
moment, due to a long moment arm which is the distance between the side and main
column (cases F, G, and K) or a column and a central intersection in case H. Cases D, E, and
I showed a moderate result, withstanding the forces with small stress concentrations in the
J. Mar. Sci. Eng. 2023,11, 1663 9 of 21
connections between structural members. Cases A, B, C, and J showed the best results due
to their ring pontoon structure.
J.Mar.Sci.Eng.2023,11,xFORPEERREVIEW9of22
Figure10.MaincolumnbendingDLP3.
3.PontoonDesign
3.1.FloatersComparison
Forthispartofthisstudy,simplifiedmodelswereused,whichhadauniformplate
thicknessof30mmandnoreinforcement,exceptforbulkheads.Theforceapplication
areaswerekeptrigidusingcoupledmulti-pointconstraints.Theadequacyofthissimpli-
ficationwasdemonstratedinapreviousstudy[9].Bycomparingthestressdistributions
ofafull-detailmodelinFigure7andthesubsequentresults,itcanbeseenthatthestress
concentrationareasarethesame(jointsbetweencolumns,pontoonsandbracings;brac-
ingsplating,thelowestbeltofcolumnplating),thusthesimplifiedmethodandmodels
producecredibleresultsforthiscomparisonstudy.
TheFEAstressdistributionscanbeseeninFigures11–15.Therearequiteafewob-
servations:
(A)Ringpontoonwithbracing (B)Triangularringpontoon
(C)TriangularRing45deg.pontoonwithbracing(D)Y-shapedpontoonwithbracing
Figure11.Cases(A–D)FEADLP1,stressplot.
Figure 11. Cases (A–D) FEA DLP 1, stress plot.
J.Mar.Sci.Eng.2023,11,xFORPEERREVIEW10of22
(E)T-shapedpontoonwithbracing(F)L-shapedpontoonwithtwobraces
(G)L-shapedpontoonwiththreebraces(H)Y-shapedbracingwithoutpontoon
Figure12.Cases(E–H)FEADLP1,stressplot.
(I)Y-shapedbracingwithsmallpontoon(J)Tubu lar joints
(K)L-shapedtallpontoonwithnobraces(L)Y-shapedpontoon,centralcolumn
Figure13.Cases(I–L)FEADLP1,stressplot.
Figure 12. Cases (E–H) FEA DLP 1, stress plot.
J. Mar. Sci. Eng. 2023,11, 1663 10 of 21
J.Mar.Sci.Eng.2023,11,xFORPEERREVIEW10of22
(E)T-shapedpontoonwithbracing(F)L-shapedpontoonwithtwobraces
(G)L-shapedpontoonwiththreebraces(H)Y-shapedbracingwithoutpontoon
Figure12.Cases(E–H)FEADLP1,stressplot.
(I)Y-shapedbracingwithsmallpontoon(J)Tubu lar joints
(K)L-shapedtallpontoonwithnobraces(L)Y-shapedpontoon,centralcolumn
Figure13.Cases(I–L)FEADLP1,stressplot.
Figure 13. Cases (I–L) FEA DLP 1, stress plot.
J. Mar. Sci. Eng. 2023, 11, x FOR PEER REVIEW 11 of 22
(A) Ring pontoon with bracing, DLP 2 (E) T-shaped pontoon, DLP 2
(F) L-shaped pontoon with 2 bracings, DLP 2 (G) L-shaped pontoon with 2 braces, DLP 2
Figure 14. Cont.
J. Mar. Sci. Eng. 2023,11, 1663 11 of 21
J. Mar. Sci. Eng. 2023, 11, x FOR PEER REVIEW 12 of 22
(K) L-shaped tall pontoon with no braces, DLP 2 (L) Y-shaped pontoon with central column,
DLP 2
Figure 14. Representative cases for DLP 2, stress plot.
(B) Ring pontoon without bracing, DLP 3 (C) Ring 45 deg. pontoon, DLP 3
(F) L-shaped pontoon with two braces, DLP 3 (H) Y-shaped bracing without pontoon, DLP 3
(J) Tubular joints, DLP 3 (K) Y-shaped tall pontoon with no bracing, DLP 3
Figure 15. Representative cases for DLP 3, pontoon stress plot.
Figure 14. Representative cases for DLP 2, stress plot.
J.Mar.Sci.Eng.2023,11,xFORPEERREVIEW11of22
(A)Ringpontoonwithbracing,DLP2(E)T-shapedpontoon,DLP2
(F)L-shapedpontoonwith2bracings,DLP2(G)L-shapedpontoonwith2braces,DLP2
(K)L-shapedtallpontoonwithnobraces,DLP2(L)Y-shapedpontoonwithcentralcolumn,
DLP2
Figure14.RepresentativecasesforDLP2,stressplot.
(B)Ringpontoonwithoutbracing,DLP3(C)Ring45deg.pontoon,DLP3
J.Mar.Sci.Eng.2023,11,xFORPEERREVIEW12of22
(F)L-shapedpontoonwithtwobraces,DLP3(H)Y-shapedbracingwithoutpontoon,DLP3
(J)Tubul arjoints,DLP3(K)Y-shapedtallpontoonwithnobracing,DLP3
Figure15.RepresentativecasesforDLP3,pontoonstressplot.
DLP1:Facedwithabeamwave-inducedtwist,floaterswithapontoonconnecting
everytwocolumnsresisttheloadwell,whilefloaterswithoutallthreepontoonsorwith
anintersectioninthemiddleshowlargeareasofstressconcentrations(casesF,G,H,and
K).Becausethestructureisnotrigidenough,asmallforceisabletogivebirthtoalarge
moment,duetoalongmomentarmwhichisthedistancebetweenthesideandmain
column(casesF,G,andK)oracolumnandacentralintersectionincaseH.CasesD,E,
andIshowedamoderateresult,withstandingtheforceswithsmallstressconcentrations
intheconnectionsbetweenstructuralmembers.CasesA,B,C,andJshowedthebestre-
sultsduetotheirringpontoonstructure.
Interestingly,caseLdidnotexhibitthesamehighstressaroundthepontooninter-
sectionasasimilarcaseH;thiscanbeaributedtoanothercolumninthemiddlerein-
forcingthestructureincaseL,aswellastheroundedcorners.However,whenfacedwith
shear(DLP2),caseLfailswith354%highermaxstresswhencomparedtocaseA,dueto
thecentralcolumnlocation.UnderthisDLP,thepontoonsconnectedinthemiddlecan
beseenasmomentarmsforlargemomentstoact.
Incaseswherethecolumnswereonlyconnectedwithpontoons,stressismuchhigher,
thanwherebothpontoons(inthebottom)andbracings(ontop)connectthecolumns.
DLP2:Generally,theconfigurationsthatwereresilienttoDLP1hadalsowithstood
DLP2well,albeitwithgenerallyhighermaximumstress,withnotableexceptionsincases
FandL.AsseeninFigure14,inthesetwocases,themomentaroundthetransverse(Y)
axiscreatedlargestressesinlongitudinalmembers,casesKandLwereabletowithstand
thatduetotheirpontoonbeingtwiceashighascomparedtootherconfigurations.This
effectisveryapparentincaseE—evenasthemaxstressdidnotgetmuchhigher,the
stressconcentrationareabecameverylargearoundthelongitudinalpontoon.Iftheplat-
formispositionedtofacethemostprobablestrongwinddirection,DLP2isthemost
probablemodeofloadingandmustbetreatedwithsufficientimportance.
Figure 15. Representative cases for DLP 3, pontoon stress plot.
J. Mar. Sci. Eng. 2023,11, 1663 12 of 21
Interestingly, case L did not exhibit the same high stress around the pontoon intersec-
tion as a similar case H; this can be attributed to another column in the middle reinforcing
the structure in case L, as well as the rounded corners. However, when faced with shear
(DLP 2), case L fails with 354% higher max stress when compared to case A, due to the
central column location. Under this DLP, the pontoons connected in the middle can be seen
as moment arms for large moments to act.
In cases where the columns were only connected with pontoons, stress is much higher,
than where both pontoons (in the bottom) and bracings (on top) connect the columns.
DLP 2
: Generally, the configurations that were resilient to DLP 1 had also withstood
DLP 2 well, albeit with generally higher maximum stress, with notable exceptions in cases
F and L. As seen in Figure 14, in these two cases, the moment around the transverse (Y)
axis created large stresses in longitudinal members, cases K and L were able to withstand
that due to their pontoon being twice as high as compared to other configurations. This
effect is very apparent in case E—even as the max stress did not get much higher, the stress
concentration area became very large around the longitudinal pontoon. If the platform is
positioned to face the most probable strong wind direction, DLP 2 is the most probable
mode of loading and must be treated with sufficient importance.
At the same time, the lack of transverse bracing had no impact on this load case,
allowing cases K and G to have lower stress than in DLP 1, which might have influenced
some real-life prototype designers motive to remove transverse bracing and pontoon.
DLP 3
: Unique for Floating Wind, DLP 3 simulates wind turbine loading detached
from other loads. Except for cases A and C, all others showed much higher stresses under
this load. As seen in Figure 15, if made without bracing, every platform is going to break.
This happens because the wave force is acting on the column top, while a pontoon
supports it only at the bottom, thus the column height becomes a huge moment arm.
In particular, all configurations with Y-shaped pontoons performed particularly bad,
with the Y-shaped bracing without underwater pontoon being the worst. This is because
of two reasons: they have a weak point at the middle intersection and they only have one
pontoon going to each of the columns, instead of two pontoons connecting each column.
At the same time, Y-pontoons use equal amount of steel to manufacture the three pontoons,
as the ring pontoon configuration. These characteristics make Y pontoons an uneconomical
option and they should be avoided.
Case C showcases how to best withstand wind turbine load—a tall and narrow pon-
toon is much better than a conventional flat and wide. Using the same amount of steel,
stress is reduced. However, such configuration’s drag coefficient might be much larger.
Case J with tubular bracing showed good results overall, but with some spiky stress con-
centrations in connection regions due to members intersecting at sharp angles. This could
be avoided if 90 degrees connections were used.
3.2. Bracing Design
Given the large stresses found, let us discuss wind turbine and bracing in detail.
During normal operation and even during typhoon conditions when the turbine blades
still receive wind loads, the turbine will generate a significant bending moment on the
supporting structure. This loading consists of three force components and three-moment
components, with the most prominent being the fore-aft bending moment and the vertical
force that includes the weight of the turbine itself. Figure 16 shows the stress distribution in
the platform, which highlights the importance of the bracing, especially in the connection
parts, as they experience the highest stresses.
As first described by Jonkman [
18
], the wind turbine installed on a FOWT can generate
several times the moment on its support that the same turbine would generate on land.
Thus, it is essential that the bracing is designed to withstand these loads to ensure the
platform can fulfil its primary purpose of supporting the turbine. Fatigue life is also
significantly reduced under these loads, further emphasizing the importance of robust
J. Mar. Sci. Eng. 2023,11, 1663 13 of 21
bracing design. Future research will investigate the impact of wind turbine moments on
bracing in more detail, and this will be the subject of an upcoming paper.
J.Mar.Sci.Eng.2023,11,xFORPEERREVIEW13of22
Atthesametime,thelackoftransversebracinghadnoimpactonthisloadcase,al-
lowingcasesKandGtohavelowerstressthaninDLP1,whichmighthaveinfluenced
somereal-lifeprototypedesignersmotivetoremovetransversebracingandpontoon.
DLP3:UniqueforFloatingWind,DLP3simulateswindturbineloadingdetached
fromotherloads.ExceptforcasesAandC,allothersshowedmuchhigherstressesunder
thisload.AsseeninFigure15,ifmadewithoutbracing,everyplatformisgoingtobreak.
Thishappensbecausethewaveforceisactingonthecolumntop,whileapontoon
supportsitonlyattheboom,thusthecolumnheightbecomesahugemomentarm.
Inparticular,allconfigurationswithY-shapedpontoonsperformedparticularlybad,
withtheY-shapedbracingwithoutunderwaterpontoonbeingtheworst.Thisisbecause
oftworeasons:theyhaveaweakpointatthemiddleintersectionandtheyonlyhaveone
pontoongoingtoeachofthecolumns,insteadoftwopontoonsconnectingeachcolumn.
Atthesametime,Y-pontoonsuseequalamountofsteeltomanufacturethethreepon-
toons,astheringpontoonconfiguration.ThesecharacteristicsmakeYpontoonsanune-
conomicaloptionandtheyshouldbeavoided.
CaseCshowcaseshowtobestwithstandwindturbineload—atallandnarrowpon-
toonismuchbeerthanaconventionalflatandwide.Usingthesameamountofsteel,
stressisreduced.However,suchconfiguration’sdragcoefficientmightbemuchlarger.
CaseJwithtubularbracingshowedgoodresultsoverall,butwithsomespikystresscon-
centrationsinconnectionregionsduetomembersintersectingatsharpangles.Thiscould
beavoidedif90degreesconnectionswereused.
3.2.BracingDesign
Giventhelargestressesfound,letusdiscusswindturbineandbracingindetail.Dur-
ingnormaloperationandevenduringtyphoonconditionswhentheturbinebladesstill
receivewindloads,theturbinewillgenerateasignificantbendingmomentonthesup-
portingstructure.Thisloadingconsistsofthreeforcecomponentsandthree-moment
components,withthemostprominentbeingthefore-aftbendingmomentandthevertical
forcethatincludestheweightoftheturbineitself.Figure16showsthestressdistribution
intheplatform,whichhighlightstheimportanceofthebracing,especiallyintheconnec-
tionparts,astheyexperiencethehigheststresses.
AsfirstdescribedbyJonkman[18],thewindturbineinstalledonaFOWTcangener-
ateseveraltimesthemomentonitssupportthatthesameturbinewouldgenerateonland.
Thus,itisessentialthatthebracingisdesignedtowithstandtheseloadstoensurethe
platformcanfulfilitsprimarypurposeofsupportingtheturbine.Fatiguelifeisalsosig-
nificantlyreducedundertheseloads,furtheremphasizingtheimportanceofrobustbrac-
ingdesign.Futureresearchwillinvestigatetheimpactofwindturbinemomentsonbrac-
inginmoredetail,andthiswillbethesubjectofanupcomingpaper.
Figure16.Stressdistributionwithsixcomponentswindloadapplied.Thegreenareaishigherstress
duetothewindturbinemoment.
Figure 16.
Stress distribution with six components wind load applied. The green area is higher stress
due to the wind turbine moment.
Note that bracing can be structurally critical. The worst structural failure in a semi-
submersible may be the Alexander Kielland accident in 1980, which resulted in the
capsizing of the floater [
19
]. The disaster started from a fatigue crack in bracing. The
damaged bracing caused the column to break off from the platform. Consequently, the
semi-submersible capsized.
A braceless V-shaped semi-sub FOWT was investigated in 2015 [
20
]. It was argued
that braceless configuration reduces fatigue stress because it removes the fatigue-prone
joints altogether, citing successful examples of braceless oil platforms. We were unable to
verify the success of such platforms and they were not referenced and most crucially, the
authors did not consider the wind turbine bending moment, which makes a huge difference
between FOWT and an oil rig. Fukushima Shinpuu was a cited example of a braceless
FOWT. This semi-sub concept was retired in 2020 due to “obvious structural failures” [
21
].
No post-decommission reports were published, but it could be deduced using this paper’s
analysis that fatigue cracks might have developed in the pontoons as they had to bear the
wind turbine moment in the absence of bracing.
Figure 17 plots a relationship between the number of connecting members and average
stress levels in the aforementioned configurations analyzed under DLP 1. High average
stress is directly related to fatigue in the case of FOWT as this stress will be caused by
highly cyclic wave and turbine loads. Connections are the sum of bracings and pontoons
of a platform. It was found that the number of pontoons does not have such a significant
impact in this case, as bracings; a trend can be seen that the fewer bracings there are, the
higher the average stress.
While it might be obvious to experienced engineers, FOWT is a new area with en-
gineers coming from different industries, who, as can be seen from many unsuccessful
designs, might be unaware of unique design requirements imposed by the wind turbine,
and this paper aims to prevent the structural failures from happening, as it might make
FOWT a less trusted technology.
Table 4shows that platforms with similar hull weights can have vastly different
structural performances. Although their average stress is not significantly different, the
maximum stress varies greatly under the same load (DLP 1). This indicates that some
shapes have severe stress concentration areas compared to others.
J. Mar. Sci. Eng. 2023,11, 1663 14 of 21
J.Mar.Sci.Eng.2023,11,xFORPEERREVIEW14of22
Notethatbracingcanbestructurallycritical.Theworststructuralfailureinasemi-
submersiblemaybetheAlexanderKiellandaccidentin1980,whichresultedinthecap-
sizingofthefloater[19].Thedisasterstartedfromafatiguecrackinbracing.Thedamaged
bracingcausedthecolumntobreakoff fromtheplatform.Consequently,thesemi-sub-
mersiblecapsized.
AbracelessV-shapedsemi-subFOWTwasinvestigatedin2015[20].Itwasargued
thatbracelessconfigurationreducesfatiguestressbecauseitremovesthefatigue-prone
jointsaltogether,citingsuccessfulexamplesofbracelessoilplatforms.Wewereunableto
verifythesuccessofsuchplatformsandtheywerenotreferencedandmostcrucially,the
authorsdidnotconsiderthewindturbinebendingmoment,whichmakesahugediffer-
encebetweenFOWTandanoilrig.FukushimaShinpuuwasacitedexampleofabraceless
FOWT.Thissemi-subconceptwasretiredin2020dueto“obviousstructuralfailures”[21].
Nopost-decommissionreportswerepublished,butitcouldbededucedusingthispaper’s
analysisthatfatiguecracksmighthavedevelopedinthepontoonsastheyhadtobearthe
windturbinemomentintheabsenceofbracing.
Figure17plotsarelationshipbetweenthenumberofconnectingmembersandaver-
agestresslevelsintheaforementionedconfigurationsanalyzedunderDLP1.Highaver-
agestressisdirectlyrelatedtofatigueinthecaseofFOWTasthisstresswillbecausedby
highlycyclicwaveandturbineloads.Connectionsarethesumofbracingsandpontoons
ofaplatform.Itwasfoundthatthenumberofpontoonsdoesnothavesuchasignificant
impactinthiscase,asbracings;atrendcanbeseenthatthefewerbracingsthereare,the
highertheaveragestress.
Figure17.Connection,bracingsandpontoonsnumberversusaveragestressatDLP1relationship.
Whileitmightbeobvioustoexperiencedengineers,FOWTisanewareawithengi-
neerscomingfromdifferentindustries,who,ascanbeseenfrommanyunsuccessfulde-
signs,mightbeunawareofuniquedesignrequirementsimposedbythewindturbine,and
thispaperaimstopreventthestructuralfailuresfromhappening,asitmightmakeFOWT
alesstrustedtechnology.
Tabl e 4showsthatplatformswithsimilarhullweightscanhavevastlydifferentstruc-
turalperformances.Althoughtheiraveragestressisnotsignificantlydifferent,themaxi-
mumstressvariesgreatlyunderthesameload(DLP1).Thisindicatesthatsomeshapes
haveseverestressconcentrationareascomparedtoothers.
R²=0.8537
0
2
4
6
8
10
12
14
0 5 10 15 20 25 30 35
Numberofelements
Averagestress,DLP1,MPa
ConnectionsvsStress Bracings
Pontoons Log.(ConnectionsvsStress)
Log.(Bracings) Poly.(Pontoons)
Figure 17. Connection, bracings and pontoons number versus average stress at DLP 1 relationship.
Table 4.
Weight (detailed model) and stress (simplified model). Data shown in per cent change from
reference case A. The darker the green, the more favorable is the result; the paler the red, the less
favorable among the same column.
Case Hull Weight,
Tones
Average Stress
DLP 1, MPa
Max. Stress
DLP 1, MPa
Max. Stress
DLP 2, MPa
Max. Stress
DLP 3, MPa
A. Ring pontoon with bracing 0% 0% 0% 0% 0%
B. Triangular ring pontoon −8% 18% 24% 34% 744%
C. Ring 45 deg., with bracing 11% −18% 26% 8% −69%
D. Y-shaped pontoon
with bracing −13% 18% 75% 21% 309%
E. T-shaped pontoon
with bracing −10% 6% 13% 11% 313%
F. L-shaped pontoon with
two braces −14% 88% 31% 0% 288%
G. L-shaped pontoon with
three braces −11% 29% 69% 13% 288%
H. Y bracing without pontoon
−19% 53% 385% 437% 2991%
I. Y bracing with
small pontoon −8% 18% 112% 140% 347%
J. Tubular joints −6% −53% −20% 16% 625%
J2 Tubular,
double-thickness tubes 53% −71% −50% −16% 278%
K. L-shaped tall, no braces −17% 88% 126% 4% 1278%
L. Y-pontoon, central column −32% 18% −52% 153% 356%
The best and worst performers in each DLP (column) are denoted by green and red
colors, respectively. Interestingly, some platforms are the best in one category but the worst
in another (cases H and L).
The advantage of configurations with many structural members is that they perform
sufficiently well under different DLPs, which is important since conditions at sea and
during transportation can be unpredictable. Configurations can withstand particular DLPs
with lower stress than full bracing and pontoon configurations, but they may fail under
other DLPs; therefore, all practically possible DLPs should be analyzed when considering
the initial structure.
J. Mar. Sci. Eng. 2023,11, 1663 15 of 21
Note that the exact value of maximum stress may differ if a high-detail model were
analyzed, but the table still provides a useful comparison of the structural performances of
different platform shapes.
The strength of a structure can be measured by its average stress. A shape called J
with tubular joints connecting columns is the strongest performer, with well-distributed
loads. However, making tubular joints is complicated because they are round and need
to connect to straight walls and each other. This causes high stress in connection regions,
but this is not a problem if the columns, pontoon, and bracing intersect at right angles and
their internal stiffeners align perfectly.
Although J is strong, it is heavier compared to other shapes with smaller pontoon and
bracing configurations. The lightest shape is H with Y-shaped bracing and no pontoon,
but it has high stress of almost 1000 MPa under DLP 1. This means it would need to be
strengthened a lot, which negates its weight economy. The shape L with a circular central
column is excluded from the weight comparison because it is a different breed of structure.
An experiment was done on J by increasing the tubular joints thickness to 60 mm,
making a new shape called J2. Its maximum stress under DLPs 1 and 2 was reduced to
below case A’s stress, and DLC 3 stress could be made equal to A’s stress with connection
reinforcement. However, J2 was heavier, weighing 6124 tons, which is 53% heavier than
shape A, which shows that just increasing thickness without changing the shape is not an
economical solution.
4. Column Design
In most of the cases presented above, the columns of the semi-submersible floater
are hexahedral, while in others they are cylindrical. This raises a question about why the
hexahedral shape was chosen. To understand this, it is important to know that the columns
of a semi-submersible floater serve three main functions:
•Supporting the wind turbine;
•Providing stability;
•Providing buoyancy.
Changing any aspect of the column to improve its performance in one function will
affect its performance in the other two functions as well. Therefore, a complex balancing
process is required to optimize the design of the columns.
At the moment, the FOWT industry has two types of column designs, as shown in
Figures 18 and 19:
Round (e.g., WindFloat designed by Principal Power) [5].
Square (e.g., Shinpuu designed by Fukushima Project) [22].
J.Mar.Sci.Eng.2023,11,xFORPEERREVIEW16of22
shapeA,whichshowsthatjustincreasingthicknesswithoutchangingtheshapeisnotan
economicalsolution.
4.ColumnDesign
Inmostofthecasespresentedabove,thecolumnsofthesemi-submersiblefloaterare
hexahedral,whileinotherstheyarecylindrical.Thisraisesaquestionaboutwhythehex-
ahedralshapewaschosen.Tounderstandthis,itisimportanttoknowthatthecolumns
ofasemi-submersiblefloaterservethreemainfunctions:
Supportingthewindturbine;
Providingstability;
Providingbuoyancy.
Changinganyaspectofthecolumntoimproveitsperformanceinonefunctionwill
affectitsperformanceintheothertwofunctionsaswell.Therefore,acomplexbalancing
processisrequiredtooptimizethedesignofthecolumns.
Atthemoment,theFOWTindustryhastwotypesofcolumndesigns,asshownin
Figures18and19:
Round(e.g.,WindFloatdesignedbyPrincipalPower)[5].
Square(e.g.,ShinpuudesignedbyFukushimaProject)[22].
Figure18.RoundcolumnsofWindFloat(picturecourtesyofPrinciplePowerInc.).
Figure19.SquarecolumnsofFukushimaShinpuu(picturecourtesyofFukushimaOffshoreWind
Consortium).
Tabl e 5liststheadvantagesanddisadvantagesofcircularandpolygonalcross-sec-
tioncolumns.Sincetheyaretwooppositeshapes,itmakessensethattheoptimaldesign
wouldbesomewhereinbetween.Tocreateacolumnthatapproximatesthecircularshape
whileonlyusingstraightplates,apolygonalcross-sectioncolumnisused.
Figure 18. Round columns of WindFloat (picture courtesy of Principle Power Inc.).
J. Mar. Sci. Eng. 2023,11, 1663 16 of 21
J.Mar.Sci.Eng.2023,11,xFORPEERREVIEW16of22
shapeA,whichshowsthatjustincreasingthicknesswithoutchangingtheshapeisnotan
economicalsolution.
4.ColumnDesign
Inmostofthecasespresentedabove,thecolumnsofthesemi-submersiblefloaterare
hexahedral,whileinotherstheyarecylindrical.Thisraisesaquestionaboutwhythehex-
ahedralshapewaschosen.Tounderstandthis,itisimportanttoknowthatthecolumns
ofasemi-submersiblefloaterservethreemainfunctions:
Supportingthewindturbine;
Providingstability;
Providingbuoyancy.
Changinganyaspectofthecolumntoimproveitsperformanceinonefunctionwill
affectitsperformanceintheothertwofunctionsaswell.Therefore,acomplexbalancing
processisrequiredtooptimizethedesignofthecolumns.
Atthemoment,theFOWTindustryhastwotypesofcolumndesigns,asshownin
Figures18and19:
Round(e.g.,WindFloatdesignedbyPrincipalPower)[5].
Square(e.g.,ShinpuudesignedbyFukushimaProject)[22].
Figure18.RoundcolumnsofWindFloat(picturecourtesyofPrinciplePowerInc.).
Figure19.SquarecolumnsofFukushimaShinpuu(picturecourtesyofFukushimaOffshoreWind
Consortium).
Tabl e 5liststheadvantagesanddisadvantagesofcircularandpolygonalcross-sec-
tioncolumns.Sincetheyaretwooppositeshapes,itmakessensethattheoptimaldesign
wouldbesomewhereinbetween.Tocreateacolumnthatapproximatesthecircularshape
whileonlyusingstraightplates,apolygonalcross-sectioncolumnisused.
Figure 19.
Square columns of Fukushima Shinpuu (picture courtesy of Fukushima Offshore
Wind Consortium).
Table 5lists the advantages and disadvantages of circular and polygonal cross-section
columns. Since they are two opposite shapes, it makes sense that the optimal design would
be somewhere in between. To create a column that approximates the circular shape while
only using straight plates, a polygonal cross-section column is used.
Table 5. Comparison of cylindrical and square columns.
Cylindrical Columns Square Columns
Pros
1. Loads are spread evenly without
stress concentrations.
2. Smaller wave impact and
drag [17].
1. Flat plates are easy to put and
weld together.
2.
Column intersects with a pontoon
at 90 degrees, making inner
structure arrangement easy.
Cons
1. Bending of plates adds time and
cost to the construction.
2. It is difficult to connect column
girders with pontoon and
bracings.
3. Increased sway and surge
amplitudes [17].
1. Potential stress concentration
points in the corners.
2. Noticeably larger drag coefficient
(not always bad).
For TaidaFloat, which has three columns and three pontoons, one column needs to
connect to two pontoons. After considering various options, a hexagonal shape is proposed
as the most suitable design. To ensure that the stiffeners of the pontoons fit the stiffeners
of the column perfectly, two of the column sides are made to be the same width as the
pontoon. Other sides are chosen to be integers and dimensions are chosen so that the main
column-to-side-column section area ratio is the same as for circular columns.
Table 6and Figure 20 provide information about the column sections.
Table 6. Section modulus and area of the columns.
Round Hexagonal
Main column
Area, m2314.16 314.38
min Wx, m3785.4 802.05
min Wy, m3785.4 753.8
Side column
Area, m2153.94 158.30
min Wx, m3269.4 275.89
min Wy, m3269.4 238.85
J. Mar. Sci. Eng. 2023,11, 1663 17 of 21
J.Mar.Sci.Eng.2023,11,xFORPEERREVIEW17of22
ForTaidaFloat,whichhasthreecolumnsandthreepontoons,onecolumnneedsto
connecttotwopontoons.Afterconsideringvariousoptions,ahexagonalshapeispro-
posedasthemostsuitabledesign.Toensurethatthestiffenersofthepontoonsfitthe
stiffenersofthecolumnperfectly,twoofthecolumnsidesaremadetobethesamewidth
asthepontoon.Othersidesarechosentobeintegersanddimensionsarechosensothat
themaincolumn-to-side-columnsectionarearatioisthesameasforcircularcolumns.
Tabl e 6andFigure20provideinformationaboutthecolumnsections.
Tab l e 5.Comparisonofcylindricalandsquarecolumns.
CylindricalColumnsSquareColumns
Pros
1. Loadsarespreadevenlywithout
stressconcentrations.
2. Smallerwaveimpactanddrag[17].
1. Flatplatesareeasytoputand
weldtogether.
2. Columnintersectswithapon-
toonat90degrees,makingin-
nerstructurearrangement
eas
y
.
Cons
1. Bendingofplatesaddstimeandcost
totheconstruction.
2. Itisdifficulttoconnectcolumngird-
erswithpontoonandbracings.
3. Increasedswayandsurgeampli-
tudes[17].
1. Potentialstressconcentration
pointsinthecorners.
2. Noticeablylargerdragcoeffi-
cient(notalwaysbad).
Tab l e 6.Sectionmodulusandareaofthecolumns.
RoundHexagonal
Maincolumn
Area,m
2
314.16314.38
minWx,m
3
785.4802.05
minWy,m
3
785.4753.8
Sidecolumn
Area,m
2
153.94158.30
minWx,m
3
269.4275.89
minWy,m
3
269.4238.85
Figure20.Circularandhexagonalsectionscomparison.TheX-axisisinred;theY-axisisinblue.
Bothtetrahedral(squaresection)andhexahedralcolumnssharethesimplicityof
manufacturingadvantage.Thereisnoequipmenttomake14m-diameterpipeswithone
seam(atleastinTai wan),socylindricalandpolyhedralcolumnsareallfabricatedby
Figure 20. Circular and hexagonal sections comparison. The X-axis is in red; the Y-axis is in blue.
Both tetrahedral (square section) and hexahedral columns share the simplicity of
manufacturing advantage. There is no equipment to make 14 m-diameter pipes with
one seam (at least in Taiwan), so cylindrical and polyhedral columns are all fabricated by
welding numerous steel plates together. In the case of polyhedral columns, steel plates can
be welded automatically into one wall section while lying flat, and then the sections are
easily assembled to form a column. In the case of cylindrical columns, every plate would
have to be bent first, requiring equipment and processing time; moreover, plates could not
be assembled into sections while lying flat because of their curvature, they would have to
be placed vertically, supported by cranes, then welded manually or semi-manually, taking
not only much more time, equipment and workers to complete the assembly, but also the
quality of manual welding is inherently inferior to automatic welding; weld strength is
particularly important for deeply submerged structures. All this equally applies to the
internal stiffeners, which would also have to be bent and welded onto the shell. Finally,
curved structures are more difficult to store and transport. Thus, it is evident that with the
same steel weight, cylindrical columns are more expensive in manufacturing, thus the exact
percentage of the cost difference is subject to local shipyard procedures and equipment.
However, the easy-to-manufacture polyhedral shapes have another perceived weakness—
stress concentrations. To check for them, finite elements analysis was conducted with
turbine weight and gravity as loads. A stress concentration was indeed found in Figure 21,
but only in corners where the bracing and pontoon connect to the column’s inner side, and
its value of 262 MPa is acceptable for the preliminary structure design stage. This corner
can easily be reinforced during subsequent design.
In Figure 22, the original TaidaFloat is shown along with two new hexagonal variants:
To determine the final shape, an international expert group consisting of twenty specialists
from the industry and academia were asked to vote. Interestingly, half of the votes went to
the original circular platform, and half went to the hexagonal platform concept.
After careful consideration, the purple hexagonal shape was chosen—as this platform
is designed to be mass-manufactured, construction ease is more important than a slightly
better wave response. As this platform is made with flat plates and the pontoon is per-
pendicular to the column side, it retains all the benefits of the square section column from
Table 1, at the same time the cons are not as bad as for the full square shape, wave impact
and drag are still less.
Following the principle of a close collaboration with the shipyard, the platform dimen-
sions were changed to fit the shipyard’s steel plate dimensions and welding equipment
sizes, e.g., pontoon and bracing height were reduced from 4 to 3.75 m to fit 3.8 m-width
steel plates produced by Taiwanese steel mills.
After all the modifications, the latest version TaidaFloat characteristics are as in Table 7.
J. Mar. Sci. Eng. 2023,11, 1663 18 of 21
J.Mar.Sci.Eng.2023,11,xFORPEERREVIEW18of22
weldingnumeroussteelplatestogether.Inthecaseofpolyhedralcolumns,steelplates
canbeweldedautomaticallyintoonewallsectionwhilelyingflat,andthenthesections
areeasilyassembledtoformacolumn.Inthecaseofcylindricalcolumns,everyplatewould
havetobebentfirst,requiringequipmentandprocessingtime;moreover,platescouldnot
beassembledintosectionswhilelyingflatbecauseoftheircurvature,theywouldhaveto
beplacedvertically,supportedbycranes,thenweldedmanuallyorsemi-manually,taking
notonlymuchmoretime,equipmentandworkerstocompletetheassembly,butalsothe
qualityofmanualweldingisinherentlyinferiortoautomaticwelding;weldstrengthispar-
ticularlyimportantfordeeplysubmergedstructures.Allthisequallyappliestotheinternal
stiffeners,whichwouldalsohavetobebentandweldedontotheshell.Finally,curvedstruc-
turesaremoredifficulttostoreandtransport.Thus,itisevidentthatwiththesamesteel
weight,cylindricalcolumnsaremoreexpensiveinmanufacturing,thustheexactpercentage
ofthecostdifferenceissubjecttolocalshipyardproceduresandequipment.
However,theeasy-to-manufacturepolyhedralshapeshaveanotherperceivedweak-
ness—stressconcentrations.Tocheckforthem,finiteelementsanalysiswasconducted
withturbineweightandgravityasloads.AstressconcentrationwasindeedfoundinFig-
ure21,butonlyincornerswherethebracingandpontoonconnecttothecolumn’sinner
side,anditsvalueof262MPaisacceptableforthepreliminarystructuredesignstage.This
cornercaneasilybereinforcedduringsubsequentdesign.
Figure21.CaseAstressdistributionunderwindturbine(deformationexaggeratedforclarity).
InFigure22,theoriginalTaidaFloatisshownalongwithtwonewhexagonalvariants:
Todeterminethefinalshape,aninternationalexpertgroupconsistingoftwentyspecial-
istsfromtheindustryandacademiawereaskedtovote.Interestingly,halfofthevotes
wenttotheoriginalcircularplatform,andhalfwenttothehexagonalplatformconcept.
Figure 21. Case A stress distribution under wind turbine (deformation exaggerated for clarity).
J.Mar.Sci.Eng.2023,11,xFORPEERREVIEW19of22
Figure22.ThreecandidatesoffloaterdesignsforTaidaFlo a t (dimensionsbeforethechangeforman-
ufacturing).
Aftercarefulconsideration,thepurplehexagonalshapewaschosen—asthisplat-
formisdesignedtobemass-manufactured,constructioneaseismoreimportantthana
slightlybeerwaveresponse.Asthisplatformismadewithflatplatesandthepontoonis
perpendiculartothecolumnside,itretainsallthebenefitsofthesquaresectioncolumn
fromTable1,atthesametimetheconsarenotasbadasforthefullsquareshape,wave
impactanddragarestillless.
Followingtheprincipleofaclosecollaborationwiththeshipyard,theplatformdi-
mensionswerechangedtofittheshipyard’ssteelplatedimensionsandweldingequip-
mentsizes,e.g.,pontoonandbracingheightwerereducedfrom4to3.75mtofit3.8m-
widthsteelplatesproducedbyTai wanesesteelmills.
Afterallthemodifications,thelatestversionTaidaFloatcharacteristicsareasinTa ble 7.
Tab l e 7.TaidaFloat’smaincharacteristics.
TaidaFloat
Ver s ion3
TaidaFloat
Ver s ion4
Displacement(t)23
,
00020
,
300
Hullweight(t)47154000
Pontoonballast(t)132409829
Columnballast(t)18503344
GM(trim/heel)(m)21.6/26.2/
13.216.8
CGheight(m)16.6217.35
Max.heelangle(°) 8.086.64
5.Discussion
Inthissection,wewilldiscussthefindingsofthisstudyandhighlightthekeyfactorsthat
needtobeconsideredwhendesigningafloatingoffshorewindturbine(FOWT)platform.
Thisstudyanalyzed13differentplatformconfigurations,underthree(3)different
designloadparameters(DLP).Theobjectivewastoidentifythemostcost-effectiveand
structurallyefficientplatformdesignthatcanwithstandtheenvironmentalloadsand
stressesexertedontheFOWTplatform.
Figure 22.
Three candidates of floater designs for TaidaFloat (dimensions before the change
for manufacturing).
Table 7. TaidaFloat’s main characteristics.
TaidaFloat
Version 3
TaidaFloat
Version 4
Displacement (t) 23,000 20,300
Hull weight (t) 4715 4000
Pontoon ballast (t) 13240 9829
Column ballast (t) 1850 3344
GM (trim/heel) (m) 21.6/ 26.2/
13.2 16.8
CG height (m) 16.62 17.35
Max. heel angle (◦) 8.08 6.64
J. Mar. Sci. Eng. 2023,11, 1663 19 of 21
5. Discussion
In this section, we will discuss the findings of this study and highlight the key factors that
need to be considered when designing a floating offshore wind turbine (FOWT) platform.
This study analyzed 13 different platform configurations, under three (3) different
design load parameters (DLP). The objective was to identify the most cost-effective and
structurally efficient platform design that can withstand the environmental loads and
stresses exerted on the FOWT platform.
Based on the analysis, platform E with a T-shaped pontoon provides the most cost-
effective solution under the assessed DLC and DLP. It is robust, relatively cheap, and easy
to fabricate. However, considering that FOWT floaters are a relatively new type of structure
with a history of failures, a safer platform configuration might be preferred for a new
15 MW
platform. In this case, the material economy is not as important as the safety of the
platform. Platforms A and C, both ring pontoons with full bracing, are considered the safest
configurations. While they are structurally similar, they may have different hydrodynamic
performance and RAO, which requires further investigation.
Additionally, it is important to consider the draft of the platform and its relation to
the pontoon. Typically, pontoons are permanently filled with water before exploitation
to counteract outside water pressure. However, in the case of a floating offshore wind
turbine, the emptied pontoon provides floatation during the transportation and installation
stages. The internal structure of the pontoon needs to be strong enough to support water
pressure exceeding 200 kPa. However, the mass of all this reinforcement helps bring
the pontoon’s center of gravity further down, improving stability. It is cheaper to add
weight by increasing pontoon steel parts thickness than by pouring concrete inside when
considering added operational costs. The added buoyancy is crucial when entering and
exiting shallow ports, allowing platforms (cases A–E) to be assembled and serviced closer
to their deployment area, while cases G–K are limited to very deep ports or have to use
temporary floatation modules.
Finally, the hexahedral column shape is chosen for TaidaFloat by the design team as
it combines the best attributes of cylindrical and square shapes. The ease of construction
with flat steel plates makes the square shape stand out.
One potential drawback of the hexahedral column shape chosen for TaidaFloat is that
it may not be as efficient in terms of space utilization compared to other shapes. Future
research could explore ways to optimize the design to maximize space utilization without
compromising on other performance metrics.
Another area for future research is the long-term durability, fatigue life and main-
tenance requirements of these designs. As FOWT floaters are relatively new structures,
there is still much to learn about their long-term behavior and performance in real-world
conditions. Continued monitoring and research can help to identify any potential issues
and optimize the design for improved durability and longevity.
There are some limitations to this study, as the method was simplified from the full-
cycle design analysis to analyze many floaters. Additionally, the shapes analyzed are
simplified, and there exist no special arrangements to make them withstand the loads better.
It would be better to design and compare the floaters in full detail including the joints and
stiffeners; however, it might take approximately a year for every floater.
6. Conclusions
In conclusion, designing an efficient and cost-effective floating offshore wind turbine
platform requires consideration of various factors such as hydrodynamic performance,
structural stability, and operational costs. Based on the research in this study, the design
team chose the following for TaidaFloat:
•
A hexahedral column, as it combines the best attributes of cylindrical and square
shapes, and uses flat steel plates, making the construction much easier and faster.
•
A full-ring pontoon configuration for low stresses under all load conditions and the
ability to control draught by ballasting.
J. Mar. Sci. Eng. 2023,11, 1663 20 of 21
•
The ring bracing configuration to prevent stress concentrations and fatigue dam-
age at connection points and efficiently distribute the wind turbine’s torque across
the structure.
The findings of this study can provide valuable insights for future FOWT platform
design and development. Further research is necessary to investigate other potential
platform shapes and configurations and optimize their performance.
Author Contributions:
Conceptualization, G.I. and I.-J.H.; Methodology, I.-J.H.; Software, G.I.;
Validation, G.I.; Formal analysis, G.I.; Investigation, G.I.; Resources, K.-T.M.; Writing—original
draft, G.I.; Writing—review and editing, I.-J.H. and K.-T.M.; Visualization, I.-J.H.; Supervision, K.-
T.M.; Funding acquisition, K.-T.M. All authors have read and agreed to the published version of
the manuscript.
Funding:
The authors highly appreciate the funding support from the government agencies, the
ROC (Taiwan) Ministry of Science and Technology, for the FOWT research as well as from Yushan
Fellow Program, Ministry of Education.
Institutional Review Board Statement: Not applicable.
Informed Consent Statement: Not applicable.
Data Availability Statement: By email.
Acknowledgments:
Many thanks to Yun Tzu Huang, who at the time worked as an engineer in
CSBC Corporation Taiwan, for suggestions on how to adjust the floater’s dimensions for optimized
production. Thanks to Jer Fang Wu, consultant and visiting scholar at NTU ESOE, for guidance on
DLP and DLC as these concepts are applied in Oil and Gas industry and ABS practice.
Conflicts of Interest: The authors declare no conflict of interest.
References
1.
Zou, J.; Udoh, I.E. Driving Down Cost: A Case Study of Floating Substructure for A 10 MW Wind. In Proceedings of the Offshore
Technology Conference, Houston, TX, USA, 6–9 May 2019.
2.
Fukushima Experimental Offshore Floating Wind Farm Project. 2015. Available online: https://www.mhi.com/news/150622en.
html (accessed on 1 August 2023).
3. Bt Floater Design. 2022. Available online: https://www.basstech.se/17/11/renewables/ (accessed on 1 August 2023).
4.
Hendriks, T.M.; Mendonca Santos, J. Challenges in Stability Assessment of offshore Floating Structures; IMAM: Lisbon, Portugal, 2017.
5.
The WindFloat
®
2021. Available online: https://www.principlepower.com/windfloat/the-windfloat-advantage (accessed on 1
August 2023).
6.
Soic Deltafloat Fowt Platform. 2021. Available online: https://www.youtube.com/watch?v=qLmMU4hBvpk (accessed on 1
August 2023).
7.
Eolfi, Pioneer of Floating Wind Power. 2022. Available online: https://www.eolfi.com/en/expertises/marine-renewable-energy
(accessed on 1 August 2023).
8.
National FOWT “San Xia Ying Ling” Design and Analysis (In Chinese). 2022. Available online: https://news.bjx.com.cn/html/
20220317/1210998.shtml (accessed on 1 August 2023).
9.
Hsu, I.-J.; Ivanov, G.E.A. Optimization of Semi-Submersible hull design for floating offshore wind turbines. In Proceedings of the
ASME 2022 41st International Conference on Ocean, Offshore and Arctic Engineering, Hamburg, Germany, 5–11 June 2022.
10.
Gaertner, E.; Rinker, J.; Sethuraman, L.; Zahle, F.; Anderson, B.; Barter, G.E.; Abbas, N.J.; Meng, F.; Bortolotti, P.; Skrzypinski, W.
Definition of the IEA 15-Megawatt Offshore Reference Wind; National Renewable Energy Lab. (NREL): Golden, CO, USA, 2020.
11.
Allen, C.; Anthony, V.; Habib, D.; Andrew, G.; Evan, G.; Nikhar, A.; Matthew, H.; Garrett, B. Definition of the UMaine VolturnUS-S
Reference Platform Developed for the IEA Wind 15-Megawatt Offshore Reference Wind Turbine; National Renewable Energy Lab. (NREL):
Golden, CO, USA, 2020.
12. Guide for Building and Classing Floating Offshore Wind Turbines; American Bureau of Shipping: Spring, TX, USA, 2020.
13.
DNVGL-ST-0119; Floating Wind Turbine Structures. CNV: Oslo, Norway, 2018. Available online: https://www.dnv.com/energy/
standards-guidelines/dnv-st-0119-floating-wind-turbine-structures.html (accessed on 1 August 2023).
14. Guidelines for Offshore Floating Wind Turbine Structures; Nippon Kaiji Kyokai: Tokyo, Japan, 2012.
15. NI-572-R02-E; Classification and Certification of Floating Offshore Wind Turbines. Bureau Veritas: Paris, France, 2015.
16.
Park, S.; Choung, J. Structural Design of the Substructure of a 10 MW Floating Offshore Wind Turbine System Using Dominant
Load Parameters. J. Mar. Sci. Eng. 2023,11, 1048. [CrossRef]
J. Mar. Sci. Eng. 2023,11, 1663 21 of 21
17.
Chen, C.-A.; Chen, K.-H.; Igarashi, Y.; Chen, D.; Ma, K.-T.; Lai, Z.-Y. Design of Mooring System for a 15mw Semi-Submersible,
Taidafloat, in Taiwan Strait. In 42nd International Conference on Ocean, Offshore and Arctic Engineering, Melbourne,
Australia, 11–16 June 2023; Available online: https://www.researchgate.net/publication/371856783_OMAE2023-104394
_DESIGN_OF_MOORING_SYSTEM_FOR_A_15MW_SEMI-SUBMERSIBLE_TAIDAFLOAT_IN_TAIWAN_STRAIT_Design_
Load_Case_EWM_Extreme_Wind_speed_Model_ESS_Extreme_Sea_State_ECM_Extreme_Current_Model_EWLR_Extr (accessed
on 1 August 2023).
18.
Jonkman, J.M. Dynamics Modeling and Loads Analysis of an Offshore Floating Wind Turbine; National Renewable Energy Laboratory:
Colorado, CO, USA, 2007.
19. Moan, T. The Alexander L. Kielland Accident; NTNU: Trondheim, Norway, 2020.
20.
Karimirad, M.; Michailides, C. V-shaped semisubmersible offshore wind turbine: An alternative concept for offshore wind
technology. Renew. Energy 2015,83, 126–143. [CrossRef]
21.
Rui Wen, S. Fukushima Energy Failed? Japan Announces Dismantling of Floating Turbines; Environment Information Center.
2020. Available online: https://e-info.org.tw/node/228593 (accessed on 1 August 2023).
22.
Ohta, M.; Komatsu, M.; Ito, H.; Kumamoto, H. Development of a V-shaped Semi-Submersible Floating Structure for 7MW
Offshore Wind Turbine. In Proceedings of the International Symposium on Marine and Offshore Renewable Energy, Tokyo,
Japan, 28–30 October 2013; Available online: http://www.fukushima-forward.jp/reference/pdf/study043.pdf (accessed on
1 August 2023).
Disclaimer/Publisher’s Note:
The statements, opinions and data contained in all publications are solely those of the individual
author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to
people or property resulting from any ideas, methods, instructions or products referred to in the content.
Content uploaded by Glib Ivanov
Author content
All content in this area was uploaded by Glib Ivanov on Aug 31, 2023
Content may be subject to copyright.
Content uploaded by Glib Ivanov
Author content
All content in this area was uploaded by Glib Ivanov on Aug 25, 2023
Content may be subject to copyright.
