ABcladapainmatrarein

KEsu IN

mvesutresumplTharThofsuAtLapaprdebual

Mfodi

C e

Tr

Dept. of C

BSTRACT: Tassic TLP (tenamage or discarticularly und time domain. aximum tensioansient effects

educed with the the 100-yr hur

EYWORDS: urvivability in m

NTRODUCTI One of the

minimal responertical modes (uitable for suppees, as well asubsea flow line

mooring systemlane motions herefore, the nre located outshe early TLP df four columnsupporting the ottached to eacater TLP desigayloads requireretension. Simesigns was acuoyancy and re., 2008), but al

The current Mexico, as docuor Design of Tistinguish betw

Corresponding e-mail: m-kim3

ansient ef

Civil Engineeri

The primary obnsion-leg platfoconnect. The derscored. The Compared to t

on on the neighs, which can lee presence of Trricane conditi

TLP tendon moderate stren

ION

main advantanses and inher(heave, pitch aporting verticas for supportines and export p

m called ‘tendonand soft f

natural frequenside the bandwdesigns for maj connected by

outer corners oh column wasgns for margined less hull b

milar intact stachieved with elatively largerlso with fewer practice for d

umented in thension Leg Pla

ween the vario

author: M. H. 3@tamu.edu

ffects of tenin moder

ing (Ocean Eng

bjective of thisform) under les

transient respnumerical sim

the common inhboring tendonead to unexpecTTRs (top-tensiion after losing

disconnectionngth hurricanes

ages of the TLrent stability, and roll), whical top-tensioneng steel catenarpipelines. The Tns’ to be very

for horizontalncies of all thewidth of incidejor hub developa ring pontoo

of a rectangulars a minimum onal fields and uoyancy and ability as formore efficien

r tendon footprtendons per co

design of TLPshe API Recomatforms (API-Rous classes of

Kim

ndon discorate-stren

Moo Hyun

gineering Prog

s paper is to iss-than-extremeponses of the pmulation is bas

dustry practicen can be significted failure ofioned risers) wg one tendon.

n; Transient ; Comparison w

LP concept is primarily in t

ch makes it med risers with dry risers to tieTLPs use verti

y stiff for vertil-plane motioe 6DOF motioent wave energpments consistn at the base ar deck at the toof three tendosmaller topsidless total tend

r the early Tt distribution

rints (e.g. Yangorner. s for the Gulf

mmended PractRP 2T), does nTLPs under t

onnectiongth hurric

Kim and Zhi

gram), Texas A

investigate thee storm conditplatform and ed on the BE-Fe of checking ticantly increasef the total syst

with pneumatic

effects; Hull-with common

its the ost dry e in cal cal ns.

ons gy. ted and op. ns. des don LP of

g et

f of ice not the

assumpfactors been rrecognistabilityamongis whetwill betotal lofailure dynamiloss oenvironthat the

Thetendon becauseincludinfailuresmodelintypicallplatformvariousto windcorrespplatformthese dperformtendons

Inter

n on the sucane cond

i Zhang

A&M University

e dynamic stabtions where onetendon tensionFE hybrid hullthe system withed at the mometem. It is also tensioners. It

-tendon-riser industry practi

ption that all are equally ap

raised as to ize what appey and survivathe various cl

ther the design e designed witoss of the pla

due to whaic survivabilityf one or tw

nment is simule incident wavee analysis of Tconditions is

e of the large ng effects of s. Another mng of highlyly in play durinm loses tendons nonlinear phed and wave l

ponding changm yaw rotationdifficulties, it m rigorous dyns in extreme

r J Nav Archit Ohttp://dx.doi.org/10.2478/IJNAOE-2013-0002

urvivabilityitions

ty College Stati

bility and survie or more tendns at the moml-tendon-riser

hout a failed teent of disconnefound that th

is also seen th

coupled dynaice

design analyspplicable to all

whether the ear to be inheability for damlasses of TLPspractice ensurh adequate rob

atform in the atever reason. y of a classic fwo tendons iated and analyes, winds, and TLP platform in general nonumber of po

inter-dependenmajor source oy nonlinear ang progressive

n restraints andenomena, such loading, varia

ge in hydrodyns etc., need tois not commonamic simulat(100-year ret

Oc Engng (200

y of a TLP

ion, TX 77843,

ivability of a fdons have beenment of disconcoupled dynamndon in the begection due to the transient eff

hat the TLP can

amic analysis

sis procedures concepts. Quedesign pract

erently differenmaged tendons. The underlyres that all clasbustness to guunlikely event

As an illusfour-column TLin a 10-year

yzed under the currents are coresponse und

ot so easy. Thiossible damagent progressive of difficulty nd transient e failure situatid progressively

as variable areable wet volumynamic coeffico be included. on for TLP dtions of the toturn period an

09) 1:13~19

P

USA

four-column n lost due to nnection are mic analysis ginning, the he snap-like fects can be nnot survive

s; Dynamic

and safety estions have tice should nt levels of

n conditions ing concern ses of TLPs

uard against t of tendon tration, the LP after the r hurricane assumption

ollinear. er damaged is is in part e scenarios, component

lies in the phenomena

ions. As the y heels over, eas exposed me and the cients, large

Because of designers to otal loss of nd beyond)

Copyright © 2009 Society of Naval Architects of Korea. Production and hosting by ELSEVIER B.V. This is an open access article under the CC BY-NC 3.0 license( http://creativecommons.org/licenses/by-nc/3.0/ ).

14 Inter J Nav Archit Oc Engng (2009) 1:13~19

environmental conditions. On the other hand, the survivability of the platform in a moderate strength hurricane with the loss of one or two tendons is still of great practical importance. In checking the case, designers remove one tendon in the beginning and run simulations. Therefore, the transient effects at the moment of disconnection are missing. To the best of authors’ knowledge, we are not aware of any prior publications (except our previous work published in a conference proceeding; Yang et al., 2008) in the public domain that have addressed dynamic survivability of TLPs with damaged tendons including transient effects. In this case, until all the tendons at one corner are totally lost, the change of heel angles and wet volumes are relatively minor and their effects can be neglected, which is assumed in the present analysis.

During the past decade, a reliable time-domain vessel-mooring-riser coupled dynamic analysis program has been developed for multi-floating systems (e.g. Ran & Kim, 1997; Kim et al., 2001) and it is applied to the present application. The hull hydrodynamic coefficients are calculated from the 3D diffraction/radiation panel program WAMIT (Lee et al., 1991). The tendon and riser dynamics are calculated by using the FE of rod equation (Garrett, 1982). The coupled hull and line dynamics are solved simultaneously in a combined matrix at every time step. The importance of the hull-mooring-riser coupled dynamic analysis for deepwater applications is well addressed in Ma et al. (2000) and Wichers et al. (2004).

Table 1 TLP Hull parameters. Description Magnitude

Draft (ft) 80.0

Total Weight (kips) 50,558

Total Tendon Pretension at the Top (kips) 15,520

Total Riser Pretension at the Top (kips) 4,348

Displacements (kips) 70,426

Vertical Center of Gravity from MWL(ft) 28.1

Vertical Center of Buoyancy from MWL(ft) -49.8

Roll Radius of Gyration (ft) 108.9

Pitch Radius of Gyration (ft) 108.9

Yaw Radius of Gyration (ft) 106.3

Wind Load Coefficient (kips/(ft/sec)2) 0.0665

Center of Pressure from MWL (ft) 125.0

Column diameter/draft (ft) 54(16m)/80(24m)

Pontoon height/width (ft) 24(7m)/27(8m)

Column center-to-center spacing (ft) 200 (61m)

TLP SYSTEM AND ENVIRONMENTS

In this study, a conventional 4-column TLP (Tension Leg Platform) with rectangular pontoons is chosen to perform the Hull/Tendon/Riser couples dynamic analysis. The hull parameters of the target TLP are listed in Table 1.

The target TLP has 8 tendons, as numbered in Fig.1, two on the outermost part of each column. There are also 8 top-tensioned risers, as shown in Fig. 1, including one drilling riser and seven production risers. In the numerical simulation, each tendon and riser is divided into 25 high-order FE elements. The properties of the tendons and risers are listed in Table 2.

Fig. 1 TLP hull, tendons, and risers layout.

Deck : 240ft × 240ft × 45ft

Porch height 7.5 ft

MWL

75ft 67ft

80ft 24ft

z

x

T1

60°

200 ft

27 ft 200 ft

54ft O. D.

TendonPorch

T7

T8

T6

T5

T4

T3

T2

15°

y

x

In

Ta

onthet str

0z0T

Fi

Pla

Te

Te

Dr

nter J Nav Arch

able 2 Material

The TTRs nboard dry-treehe numerical m

al. (2007). Troke z of the te0 = effective le

0= initial top te

ig. 2 TLP Num

Wet Weight

Dry Weight

Diameter

EA

EI (kips*ft2) Inertial

Coefficient

Drag Coeffici

Top Pre-tension

Break Streng

Production Ris

atform Hull

endon #3&#4

endon #5&#6

rilling Riser

hit Oc Engng (2

l properties of

(top-tensionede facilities by

model of which The relation bensioner is givength of gas iension at strok

merical model.

t

t

t

ent

n

gth

ers z

(2009) 1:13~19

the tendons an

d risers) are ca flexible pneuis given, for e

etween the teven by 0 /T T=in the associa

ke z = 0, and n =

Tendons

68.98 lb/ft (102.65 kg/m)

425.66 lb/ft (634.84 kg/m )

2.67 ft (0.81 m)

3.76×106 kips (1.67×107 kN)

3.08×106 kips*f(1.27×106 kN*m

2

1

1.94×103 kips (8.63×103 kN)

7.59×103 kips (3.38×104 kN )

x

z

y

9

nd risers.

connected to tumatic tensionexample, in Yansion T and t

0(1 / )nz z+ , wheated accumulat= gas constant

)

) ft2

m2)

)

)

Truss Member

Tendon #1&#

Plate Membe

Tendon #7&#

the

ner, ang the ere tor, .

Theplatformpanel ethe hydhull is truss mrespectfluctuatformula

Themoorinorder suthey aranalysiexcitatiapproxi

Theranges 0.05 semuch aappliedperiod o

Twand curyear huTableirreguladynamiwith piincideneffects designetendon,intentiosimulat

Drilling

173.68 (258.86 k

327.3 l(488.19 k

1.75 (0.53

2.21×10(9.83×100.00 kip(0.00 kN

2

1

7.29×10(3.24×10

2.85×10(1.27×10

rs

#2

rs

#8

e water depth m hull is 1.57elements for a drodynamic commodeled by u

members. The vive hull elemetions are calcua e panel mod

ng/riser arrangeum-frequency re not consids. However, tions are incluimation methoe time-domainof 2500 ~ 50

econds. To supas possible, th

d from zero toof 250 seconds

wo environmentrrent considereurricane condit3 and 4). JOar-wave generic wind-velociiecewise-linearnt angles (0 deg

of environmened to survive , the weathonally forced ttion to obser

Riser

lb/ft kg/m )

lb/ft kg/m )

ft m) 6 kips

06 kN ) ps*ft2 N*m2)

2 kips 03 kN )

3 kips 04 kN )

is 3000 ft an× 106 lb (712quarter of the mputation. The

using 4 equivalviscous forces

ents including ulated through

del of the ement are showwave excitatio

dered to be imthe second-orduded through d.

n simulations 000 seconds. Tppress the starthe environmeno the actual vs. tal conditions wed here, are thtion in the GOONSWAP wavration and APity generationr decay rates igree and 45 degntal heading anfor 10-year st

her-side (taut-to break at therve the subse

Pr

nd the total m2,800 kg). A to

platform hull e damping of tlent plate mems and the dampthe effects of

h the modified

platform hulwn in Fig. 2. Tons are not incmportant for der slowly vathe so-called

were carried The time step t-up transient rntal force wa

value during t

with collinear whe API 10-yea

OM (Gulf of Mve spectra arPI spectra ar. A steady shis imposed. Twgree) are chosengles. Since thorm with the -side) tendone bottom porchequent behavi

roduction Riser

115.41 lb/ft(172.24 kg/m )

157.56 lb/ft (234.91 kg/m )

0.92 ft (0.28 m)

9.75×106 kips(4.34×106 kN)0.00 kips*ft2

(0.00 kN*m2)

2

1

5.17×102 kips(2.30×103 kN )

2.02×103 kips(8.99×103 kN )

15

mass of the otal of 1420

are used in the platform mbers and 4 ping on the free-surface

d Morison’s

l and the The second-cluded since the present

arying wave Newman’s

out in the is set to be

responses as as gradually the ramping

wind, wave, ar and 100-

Mexico) (see re used for re used for hear current wo different en to see the he TLP was loss of one

n #5 was h during the iors of the

rs

16

pldiwdifrenobe

Ta

Ta10

N

dedubutraTodaintimposimha

thThsim

6

latform includisconnection. Taves reaches tisconnection, tee-hanging. Hoot supposed toe shown in the

able 3 Wave an

Significant W

Wave Pea

Enhancement

One Hour MVelocity at 10

Current Surfa

able 4 Current00 years condit

Depth

Surfac

-165.5 ft (-5

-331.0 ft (-10

-1000 ft (-304

NUMERICAL When a tend

efect in fabricauring service ourden of extra bansferred to ano investigate thamaged conditntentionally disme when a hosition of the Tmulation until appens at the b

We first conhe TLP is supphe supposed wmulated 6-DOF

ding transientThe breaking tithe instantaneothe failed tendowever, in the

o survive after next section.

nd wind condit

Wave Height

ak Period

Parameter (ã)

Mean Wind 0m Elevation

face Velocity

t conditions intion.

h

e

0.4 m)

00.9 m)

48.0 m)

L RESULTS A

don is removedation, damage or accidental unbuoyancy and

nd redistributedhe survivabilitytion in a mo

sconnect the tahigh wave peTLP (462.5s) 2500s. It is a

ottom porch. nsider the caseposed to surviv

worst heading, 4F motions are p

t effects at ime is set wheous position odon is still atta

100-year condlosing one ten

tions.

10-yr

31.8 ft (9.7 m)

12.7 sec

2.4

86.0 ft/sec(26.2 m/sec

7.9 ft/sec (2.4 m/sec)

n both 10 yea

Current V

7.9 ft/sec (2

5.9 ft/sec (

0.0 ft/sec (0

0.0 ft/sec (0

AND DISCUS

d from the systduring tow, crnlatch from bostabilizing righ

d among the remy of the TLP u

oderate-strengthaut-side tendoneak reaches thand continued

assumed that t

e of 10-year huve after the los45 degrees, is spresented in Fi

the moment en a peak of hif the TLP. Afached to the hdition, the TLPndon, which w

100-yr

51.8 ft(15.8 m)

15.4 sec

2.4

c c)

144.8 ft/se(44.1 m/se

)

ars condition a

Velocity

2.4 m/sec)

1.8 m/sec)

0.0 m/sec)

0.0 m/sec)

SSIONS

tem (either duerack developmeottom porch), thting momentsmaining tendounder the tendoh hurricane, n #5 at a specihe instantaneo

d to carry out tthe disconnecti

urricane in whiss of one tendoselected first. Tig. 3.

of igh fter hull P is will

ec ec)

and

e to ent the s is ns. on-we ific ous the ion

ich on.

The Fig. 3and wittendon

pitc

h di

s (d

eg)

4

2

-2

-

-

s

urge

dis

(ft)

r

oll d

is (d

eg)

heav

e di

s (ft

)

yaw

dis

(deg

)

-

-

0

-0

s

way

dis

(ft)

Inter

TLP-motion tth risers API #5 forced to br

0 500

0 50

0 50

0 500

0 500

0 50

400

200

0

00

0.5

0

-0.5

-1

0.1

0

-0.1

400

200

0

-200

10

0

-10

-20

1

0.5

0

0.5

J Nav Archit O

time series co10 year’s condreak at 462.5 s

0 1000

00 1000

00 1000

0 1000

0 1000

00 1000

Time(

Time(s

Time(s

Time(

Time(s

Time(

Oc Engng (200

omparison, modition, 45° incsec.

without risers

with risers

1500 20

1500 20

1500 20

1500 20

1500 20

1500 20

(s)

s)

s)

(s)

s)

(s)

09) 1:13~19

odel without ident angle,

00 2500

000 2500

000 2500

00 2500

00 2500

000 2500

In

Fiduin

s

urge

dis

(ft)

hea

ve d

is (f

t)

p

itch

dis

(deg

) sw

aydi

s(ft

)ro

lldi

s(d

eg)

yaw

dis

(deg

)nter J Nav Arch

ig. 4 TLP-motiuring and befoncident angle, te

g(

)(

)p

(g)

sw

ay d

is (f

t)

rol

l dis

(deg

)

y

aw d

is (d

eg)

0

0

0

0

0

0

300

200

100

0

300

200

100

0

10

0

-10

-20

1

0.5

0

0.5

0.5

0

-0.5

-1

0.1

0

-0.1

hit Oc Engng (2

ion time seriesore simulation,endon #5 force

Tendon

Tendon

500 1000

500 1000

500 1000

500 1000

500 1000

500 1000

Tim

Tim

Tim

Tim

Tim

Tim

(2009) 1:13~19

s comparison, t, API 10 yeared to break at 4

n #5 broken in sim

n #5 broken at init

0 1500

0 1500

1500

1500

1500

1500

me(s)

me(s)

me(s)

me(s)

me(s)

me(s)

9

tendon #5 brok’s condition, 4462.5 sec.

mulation

tial

2000 2500

2000 2500

2000 2500

2000 2500

2000 2500

2000 2500

ken 45°

In tset, whImmedpitch-rointervaltendon withoutsurvivamotionregion transienlittle stitop-ten

In cases inand ducases, rwhen crunningin the domaineffects the respredictecases arthe disc

Fig. 5 Ttransien(model

0

0

0

8

6

4

2

0

8 6 4 2 0

8 6 4 2 0

Top

the horizontal hich results in iately after theoll (about 0.5-0l due to transie

tension. We t risers to see t

ability and trans are in genein pitch/roll,

nt overshoot. Aiffer both in hosioned risers. the next figurn which tendo

uring simulatiorisers are not ihecking the su

g the dynamic beginning, wh

n analysis. In at the momensulting maximed. Fig. 4 showre almost idenconnecting tim

Top Tension tint effects, APIwithout risers

×106

0 500

0 500

MBL

MBL

MBL

0 500

×106

×106

p tension (lb)

plane, the TLPabout 14-ft (

e taut-side tend0.6 degrees) oent effects, wh

compared ththe effects of t

nsient responseeral small exc

where the riAs expected, thorizontal and v re, Fig. 4, weon #5 is discoon (at 462.5sencluded. The c

urvivability witsimulation wit

hich can also this approach

nt of disconnecmum tendon-tws that the rotatical after the l

me, where the tr

ime series comI 10 year’s con).

Time(s)

1000

1000

1000

)

Time(s)

Time(s)

P experiences (4 m) set dowdon is lost, theccurs in a veryhich will in turhe cases with the riser systemes. The differencept the narroisers tend to he TLP systemvertical directio

e compared twonnected in theec), respectivecommon industh the loss of onthout includingbe treated by

h, however, thction are not intension may ational motionloss of tendon ransient oversh

mparison with andition, 45° inc

1500 2000

1500 2000

Tendon #5 breaTendon #5 brea

1500 2000

17

a large off-wn in heave.

e maximum y short time rn influence

risers and m on TLP’s nces in hull

ow transient reduce the

m becomes a ons with the

wo different e beginning

ely. In both stry practice ne tendon is g the tendon frequency-he transient ncluded and

be under-s of the two except near

hoot occurs.

and without cident angle,

0 2500

0 2500

ak duringak at

0 2500

18 Inter J Nav Archit Oc Engng (2009) 1:13~19

The corresponding time histories of tensions for both cases on tendon #6(taut-side), #4(lateral), and #2(lee-side) are also given in Fig.5. After the loss of one tendon, the mean tension on the remaining tendons is suddenly increased to counter-balance the total net buoyancy. The mean tension of the lee-side tendon is, however, decreased due to the heel angle after damage. We can also see the pronounced increase of maximum tension on the neighbouring #6 tendon at the moment of disconnection due to snap-like transient effects (Maximum tension at the top node equals to 7.22 × 106 lb with transient effects and 6.52 × 106 lb without transient effects). If the tendon is to break at the 90% of MBL, which is actually so in many real cases, the #6 tendon also fails due to the transient effects, which cannot be detected by the alternative approach without the transient overshoot. When compared with the tension on #4-tendon, it is seen that the transient effects are the most important to the neighbouring tendons. It is interesting that the tension on the lee-side tendon (#2) can suddenly decrease due to the sudden downward pitch-roll angles (see Fig.3) that may cause unwanted transient compression(or buckling) loading there.

Fig. 6 Tendons’ top tension time series comparison between models without and with risers, API 10-year condition, 45° incident angle, tendon #5 forced to break at 462.5 sec.

Fig. 6 shows the comparison of maximum transient

tension on the adjacent #6-tendon with and without risers. (Maximum tension at the top node equals to 7.22 × 106 lb without risers and 6.62 × 106 lb with risers). It is seen that the risers, by adding additional stiffness and resistance, tend to reduce the maximum transient tension, so help the survivability of the system by sharing the burden. Fig. 7 shows the time history of #6-production-riser tension. The maximum top tension of the risers reached 7.39×105 lb due to the same transient effects. The present pneumatic tensioners function positively to reduce the sudden increase of tendon-tension.

Fig. 8 shows the tension of the neighboring #6-tendon and sideway #4-tendon under the same #5-tendon-loss scenario but the environment is changed to 100-year condition.

It is seen that the neighboring tendon also breaks immediately after the disconnection and the burden is transferred to the remaining tendons to progressively fail. It needs to be reminded that the TLP is not so designed as to survive in the 100-year condition after the loss of one tendon. It is also found from our simulation that after losing two tendons in the taut-side corner, the pitch/roll angles are suddenly increased to 25°, so there should be sudden change in wind, wave, and current loadings and hydrodynamic coefficients after that. Therefore, a special care is needed to continue the simulation.

Fig. 7 #6 Risers’ top tension time series, API-10 year condition, 45° incident angle, tendon #5 forced to break at 462.5 sec (model with risers).

Fig. 8 Top-tension time series, model without risers, API 100-year condition, 45° incident angle, tendon #5 forced to break at 473.0sec.

0 500 1000 1500 2000 25001

2

3

4

5

6

7

8x 106 Tendon #6

Top

tens

ion

(lb)

time(s)

Model without risersModel with risersMBL

0 500 1000 1500 2000 25005

5.5

6

6.5

7

7.5 x 105 Riser #6

Top

tens

ion

(lb)

time(s)

Top Tension Time SeriesMean Tension Before BreakMean Tension After Break

0 500 1000 1500 2000 25000

2

4

6

8x 106 Tendon #4

Top

tens

ion

(lb)

time(s)

0 500 1000 1500 2000 25000

2

4

6

8x 106 Tendon #6

Top

tens

ion

(lb)

time(s)

Top TensionMBL

7.5

7

6.5 6

5.5

5 0 500 1000 1500 2000 2500

×106

0 500 1000 1500 2000 2500 Time(s)

Mean tension after break

MBL

MBL

Mean tension before break

Time(s)

Tendon #6

Tendon #4

Riser #6

Tendon #6

Model without risers Model with risers

MBL

Top

ten

sion

(lb

) To

p t

ensi

on

(lb)

Top

ten

sion

(lb

) Top

ten

sion

(lb

)

Time(s)

0 500 1000 1500 2000 2500

×106

8

7

6

5

4 3

2

1

0 500 1000 1500 2000 2500

0 500 1000 1500 2000 2500Time(s)

×106

×106

In

dethtencatwcotenco

Fiin

C

clamirrinvansodylinsyretenunsimmwhtratoovchextra

Top

ten

sion

(lb)

nter J Nav Arch

So far the eegrees, which ihe tension timendon in 10-yea

ase (without riswo. Therefore,ompared to thension on the n

ompared to 7.2

ig. 9 Top-tensncident angle, te

ONCLUDIN

The dynamicassic TLP(tensi

moderate-strengthregular waves,

nvestigated by nalysis programolved by BE ynamics are sone dynamics aystem in a comesponses of thnsions at the nderscored. Comulation with

maximum tensiohen the neighansient effectstal system. T

vershoot effechecking the suxtreme environansient effects

MBL

Mean tension after break

×106

p(

)

0 500

8

7 6

5

4

3

2

1

hit Oc Engng (2

environmental is supposed to history of thear hurricane cosers), the taut-s, the transfer

e 45° heading cneighboring ten2 × 106 lb in 4

sion time seriendon #5 force

G REMARK

c stability and ion-leg platformth collinear hur

dynamic windusing the hul

m in time domai(Boundary E

olved by FE mare solved sim

mbined matrix ahe platform a

moment of ompared to thehout a failed on on a tendonhboring tendon, which can leTherefore, thects in tendonurvivability ofnmental condits can be reduc

Mean tension before break

Tend

0 1000

Tim

(2009) 1:13~19

heading has bbe the worst ca #6-tendon aft

ondition with 0side tendons arred burden iscase (see Fig.5ndon in this ca5° case.

es, API 10-yeed to break at 4

KS

survivability om) after the lossrricane conditiods and steady ll-tendon-riser in. The hull hylement) methomethod. The hmultaneously at each time stand the corresdisconnection

e common indutendon in the

n can be signifn breaks due ead to unexpece inclusion on tension is f a TLP systetions. It is als

ced with the pr

--------- Top tensio

on #6

1500

me(s)

9

been fixed at ase. Fig. 9 shoer the loss of #° heading. In tre four insteads relatively le

5). The maximuase is 6.86×106

ear condition, 462.5 sec.

of a four-colums of one tendon

ons (unidirectioshear currents)coupled dynam

ydrodynamics aod and the lihull motions aas an integrattep. The transiesponding tend

are particulaustry practice ie beginning, tficantly increasto the snap-lited failure of t

of the transieimportant whem in less-thaso found that tresence of TT

on time series

2000 25

45 ows #5-his

d of ess um 6 lb

0°

mn n in nal ) is mic are ine and ted ent

don arly i.e. the sed ike the

ent-hen an-the

TRs

(top-tenadditionangle iheading100-yeathe tranunder tcorner, dynamieffects

ACKN

Thi(Mineraappreci REFER API, 2

legAPI, 20

GuGarrett

EnKim, M

TLme

Kim, MfreGa

Lee, C.ThOffSta

Ma, Wno

Ran, Z.tet7(2

WicheronOT

Yang, CandmoOM

Yang, C20gloCoPo

500

nsioned risers)nal resistance is found to beg. It is also sear hurricane cnsferred burdenthe extreme co

the heel angic analysis shwhen necessar

NOWLEDGEM

is research wals Managemeiated.

RENCE

007a. Planning platforms. Dr007b. Interim gulf of Mexico. B, D.L. 1982.

nergy Res TechM.H. Tahar, ALP motion ethodologies /pM.H. and Yueequency waveaussian seas. Jo.H. Newman, J

he computationffshore Mechavanger, Norw., Lee, M.Y., Znlinear couple. and Kim, M.thered spar in w2), pp.111-118rs, J. and Devn the deepstar tTC #16582. Chan K. Tahard nonlinear aodelling for spMAE ’07, 10-15Chan K. Padm08. Transient obal motion of onference onortugal.

since the pneand stiffness

e worse for sueen that the Tcondition aftern is too much ondition. After gle can be lahould includery.

MENT

was financiallent Service) an

ng, designing, raft RP2T, Revguidance on huBulletin, 2INT

Dynamic anahnology , 104, pA. and Kim, Y

analysis agparameters. Proe, D.K.P., 199e loads on a ournal of Ship J.N. Kim, M.Hn of second-ordanics and A

way Zou, J., and Hud analysis toolH., 1997. Nonwaves. Journa. lin, P.V. 2004theme structur

r, Arcandra andapproach of hypar global perf5 June 2007, S

manabhan, B. Meffect of tendETLP. ProceeOMAE, 15-2

umatic tension. The 45-degrurvivability tha

TLP cannot surr losing one tefor the remainlosing all ten

arge, and the e such highly

ly supported nd the funding

and constructv.3. urricane condi-MET. alysis of slendpp.302-307. Y.B., 2001. Vagainst variouoc. Int. Offshor91. Sum- and

body in uniResearch, 35,

H. and Yue, Dder wave loadsArtic Eng. C

uang, E., 2000l. Proc. OTC #

nliear coupled aal of Offsore &

4. Benchmark res FPSO, spar

d Kim, M.H. 2ydro-pneumatiformance. Pro

San Diego, USAMurray, J. and don disconnecedings of 27th In20 June 200

19

ners provide ree incident an 0-degree rvive in the endon since

ning tendons ndons at one

subsequent y nonlinear

by MMS g is greatly

ting tension

itions in the

der rods. J

ariability of us design re

difference-i-directional pp.127-140 .K.P., 1991. s. Proc. 10th Conference,

. Deepwater 12085. analysis of a

& Polar Eng.

model tests r, and TLP.

2007. Linear ic tensioner

oceedings of A . Kim, M.H.,

ction on the nternational 08, Estoril,