W

UGMSeptDec2008

Orcina
www.orcina.com

UGM2008FatigueAnalysis 1. Introduction
Thispaperpresentsacasestudyfatigueanalysis.Theanalysiswaspartofthe testingofthenewfatiguefacilitiesinOrcaFlex9.2.Therearequiteafewobjectives tocover: Tocomparelifetimeestimatesfromirregularwaveandregularwavefatigue analyses ToidentifyefficientwaysofusingOrcaFlexforfatiguecalculations Toimproveourunderstandingofthelimitsofbothirregularandregular waveanalysis Thefirstconclusiondrawnisthatthedifferentfatiguemethodsproducedifferent resultsforthesystemconsideredhere.Regularwaveanalysisisfoundtobemore conservative,andthisisinvestigatedinmoredetail. Whenworkingwitheithermethod,performingtwofatigueanalysesis recommended.Thefirstanalysisisessentiallyascreeningoftheloadcases,whichis intendedtorunasquicklyaspossible,andprovidesinformationthatallowsthe secondanalysistoproduceaccurateresultswiththeminimumofeffort.This approachisbasedontheresultsfromoursensitivitystudiesonfatigueinput. Thisdocumentbeginswithadescriptionofthesystemstudied,andthenpresents theworkperformedusingirregularwaves.Anaccuratevalueofthesystemsfatigue lifeisintroducedassoonaspossible,andweusethatfigureasareferencefor sensitivitystudiesandcomparisontoregularwaveanalysis.

2.

CaseStudySystem

Figure21showsthesystemconsideredforfatigueanalysis;a9steelcatenary riser,1600mlongin1000mwaterdepth. Theriserishungfromthepontoonofasemisubmersibleplatform.The environmentalloadsappliedtothesystemcomprise30storms,representinga condensedGulfofMexicoirregularwavescattertable.Weareawarethatusing30 casesisquiteasmallsetforafatigueanalysis,butcondensingthestormscatter casesallowednumerousrepeatanalysesinareasonableamountoftime. Aflexjointisusedattheconnectionoftherisertotheplatform,withastiffnessof 10kN.m/deg.Risertouchdownisatapproximately1175marclength. Thefatigueloadcasesareallappliedinthefardirection,whichisintendedto concentratedamageintheplaneoftherisercatenary.

UGM2008FatigueReport.doc

1of10

18December2008

UGMSeptDec2008

Z X

Orcina
www.orcina.com

Orc aFlex 9.3a2: SCR bas e c as e.dat (modified 15:16 on 09/10/2008 by Orc aFlex 9.3a2) (az imuth=270; elev ation=0) Static s Complete

200 m

Z X

Figure21Casestudysteelcatenaryrisersystem.Theriserisattachedattheoutsideofthe pontoon

3.

IrregularWaveAnalysis

Thefirstanalysisperformedusesstandardirregularwavesimulations.The30 stormcasesareeachrunforonehourduration.Servicelifeistakenas20years,and soweaimforaminimumof200yearsfatiguelife. Theminimumfatiguelifeisobtainedinthetouchdownregion,andis236years. Havingfoundaresultwebelievetobereliable,wecanworkonwaystomakethis analysismoreefficient. Irregularwaveanalysisiscomputationallydemanding.Onehourdurationstorm casesrequirequitelongruntimes,andgeneratelargesimulationfiles,requiring largeamountsofstorage.Finally,thepostprocessingoftheselongtimehistories canalsobeslow. TheOrcaFlexfatiguemodulenowcalculatesdamagebasedonlyonZZstressranges atthepipesouterstressdiameter,whichhasmadethefatiguepostprocessing muchfaster.OrcaFlexalsonowmakesmoreefficientuseofcomputerswithmore thanoneprocessor,extractingfatigueresultsfrommultiplesimulations simultaneously(processingasmanyfilesatonceasthereareavailableprocessors). Theseimprovementstoprocessingspeedhavehelpedtomakeirregularwave fatiguemuchfaster,butwedecidedtolookattheanalysis,totryandidentifyifthe processcouldbemademoreefficient. Stormdurationhasadirectimpactonthetimetakenforirregularfatigueanalysis. Individualsimulationswilltakelesstimetorunshorterstorms,andthesmallerfiles producedcanbepostprocessedmorequickly.ThisisconsideredindetailinSection 3.1.

UGM2008FatigueReport.doc

2of10

18December2008

UGMSeptDec2008

Orcina
www.orcina.com

3.1. StormDurationSensitivity
Whenusingirregularwaves,itisimportantthatthestormdurationislongenough togenerateastatisticallyacceptableseastate;onewheretheaverageheightand periodfromtherealisationareclosetotheinputvalues.Itisalsoimportantthatthe fatiguedamagecalculatedfromthesimulationsconvergestoareliablevalue. Thiscanbetestedbyrepeatingananalysiswithadifferentrandomwaveseed. Changesofseedgivedifferentwavetrainrealisations,butthefinalfatigueresults shouldnotbemuchaffected,becausethestormswillbestatisticallysimilar. Webeganbyrepeatingtheonehourfatigueanalysis,using30differentwaveseeds. Theresultsexhibitedspreadof8%aroundouracceptedtruevalue.Thisis acceptablysmall,andweconcludethatonehourstormsarelongenoughfor accurateanswers. Butwouldshorterstormsalsogiveaccurateresults?Toinvestigatethepossibility, weranfurtherfatigueanalysesusingstormdurationsof40,20,10and5minutes. Eachanalysiswasrepeated30times,usingdifferentwaveseeds. Thespreadofthefatigueresultsaroundourtruevalueatthedifferentstorm durationsisshowninFigure31.

VariationofFatigueLifewithStormDuration 450 400 350 300 250 200 150 100 50 0 0 5 10 15 20 25 30 35 40 45 50 55 60 65 StormDuration(mins)

Figure31Variationinfatigueresultsduetochangingwaveseedforgivenstormdurations

Eachpointaboverepresentsafullfatigueanalysis.Ourtruevalue(themeanofthe analysesusingonehourstorms)isshownasagreenline.
UGM2008FatigueReport.doc 3of10 18December2008

FatigueLife(years)

UGMSeptDec2008

Orcina
www.orcina.com

Thestandarddeviationoftheresultsaboutthistruevaluevarieswiththesquare rootofstormduration,T.Thismeansthathalvingthestormdurationdoesnot doublethedeviationofthepointsabove.Theincreaseinspreadislesssevere. Using20minutestormdurations,thespreadwas14%forthissystem.Webelieve thatinmanycasesthislevelofvariationmayprovideanacceptableaccuracyforan initialassessmentoffatigueperformancedependinghowclosetheresultsareto acceptablevalues. Toclarifythispoint,ifasystemrequiresaminimum200yearfatiguelifeandan initialassessmentusing20minutestormdurationspredictsa300yearfatiguelife, thenevenassumingaspreadof30%aboutthetruevalue,theresultsareacceptable. Thiskindofassessmentwillbeusefulearlyinthedesignprocess,wherearapid evaluationofperformanceisneeded.Further,whenamoreaccuratefatigueanalysis isrequired,thispreliminaryworkwillallowamoreefficientapproachtothe repeatedcalculations. Fatiguedamageisrarelyevenlydistributedoverthesetoffatigueloadcases.Some ofthestormsrunwillcontributeverylittletothetotaldamage,whileotherswill takeupalargeproportionofthefatiguelife.Clearlyagoodqualityseastate(along duration)isrequiredforthecasesthatprovidemostofthedamage,buttheless criticalcaseswilltolerateashorterstormduration. Toillustratethispoint,Table31showsthepercentagedamagecontributionfrom eachofour30irregularloadcases,laidoutonascattertablejustasenvironmental dataispresented.Thedataweretakenfromoneofthefatigueanalysesperformed usingonehourdurationstorms.Increasingcontributionsofdamagearehighlighted, withredseastatesconsumingthemostfatiguelife.

UGM2008FatigueReport.doc

4of10

18December2008

W

Hs / Tp 0.3 0.8 0.9 1.5 0.9 2 2.7 2 3.2 4.6 6.4 1.8 3.7 5.1 6.8 8.7 10.6 1.9 3.5 5.2

UGMSeptDec2008

7 9 11

Orcina
www.orcina.com

1 0.00%

3 0.00%

13

15

0.00% 0.00% 0.00%

0.00% 0.00% 1.29% 1.29% 1.29%

0.56% 0.56% 35.35% 7.03% 0.26% 0.06% 2.74% 27.88% 15.40% 5.12% 0.26% 0.00% 0.29% 0.63%

Table31TotalDamagecontributedfromeachoftheirregularloadcasesconsidered

Thisinformationwhichcasescontributethemajorityoffatiguedamageis providedbytheinitialanalysisevenwhenshortstormsareused.Thesecond analysiscantakeadvantageofthisknowledgebyusinglongstormdurationsfor criticalcases,andshorterstormsforlessimportantloadcases.The20minuteruns couldbereusedforcasescontributinglittledamage.

4.

RegularWaveAnalysis

Regularwaveanalysisprovidesasimplerapproachtofatiguecalculations.Irregular wavesprovidearealistichistoryofload,whichisanalysedbyrainflowcountingto calculatedamageoveragivenexposure. Regularwavesareassumedtogeneratethesamestressrangeeachcycle,andthis loadingisappliedtothesystemforthenumberofwavecyclesinthetotalexposure. Tousethissimpleassumption,irregularwaveseastatesareexpandedintotheir individualwaves.Thesesinglewavesareallocatedtobinsofgivenheightand period,andthereforeeachbinaccumulatesanumberofcyclesforachosen exposure.Eachbinisrepresentedbyasingleregularwavesimulationof appropriateheightandperiod.Thesesimulationsareofshortduration,and consequentlyrunquickly.
UGM2008FatigueReport.doc 5of10 18December2008

UGMSeptDec2008

Orcina
www.orcina.com

4.1. OrcaFlexWaveScatterConversion
Theconversionofirregularwaveloadcasestoasetofregularwaveswith associatednumbersofcyclescanbedoneusingOrcaFlex.TheWaveScatter Conversiontooltakesasetofirregularwaveloadcasesasinput,andwillproduce regularwavecaseswithamatchingtotaltimeofexposure.Thecyclesofeach regularwavearecalculatedusingthejointprobabilitymethodofLonguetHiggins, referencedintheOrcaFlexhelp. Oncetheirregularwaveloadcasesareenteredintotheconversionmodule, OrcaFlexwilloutputatableofregularwaveoccurrences,andthencanconvertthis tableintoanOrcaFlexbatchscript,andacorrespondingfatigueanalysisfile,with thedataonsimulationfilenamesandassociatedcyclesenteredautomatically.This functionalityisalsoavailableifasetofregularwaveloadcasesissuppliedtothe wavescatterconversiontoolnoconversionisperformed,buttheautomationcan stillbeused. Thisfacilitymakessettinguparegularwavefatigueanalysisveryeasy.Regular waveanalysisisalsolesscomputationallydemandingthanirregularwave calculations.Thesimulationfilesaresmaller,runmorequickly,typicallyinclude only5waveperiodsduration,andnorainflowcountingisrequiredontheoutput. Thestressrange(frommaximumandminimumvalues)istakenoverthelastwave cycleofsimulation. Theconversionofirregularseastatesintoregularwavesinvolvesadiscretisationof acontinuousspectrumintobins.Thismeansthatusingalargenumberofsmallbins overafixedrangeofheightandperiodwillgiveamoreaccuraterepresentationof realconditions.Acoarsediscretisation,usingfewerbinsoverthesameheightand periodrange,willbefurtherfromrealconditions. Foranydiscretisation,itisbelievedthattheprocessofconversionfromirregular stormcasestoregularwaveoccurrencesislikelytoincreasethenumberoflarge wavesappliedtothesystem.Wheneachirregularstormisconsideredbythe conversionprocess,itisassumedthatthestormpersistsforadurationequaltoits statedexposure. Forexample,theirregularwavecaseusingHs=5.1mandTp=11shasanexposure of628hours,alittleover26days.Thisisfarlongerthanasingleseastateistypically assumedtoprevail,butweconsideraseastateofthesepropertiesforthisduration. Clearlythisallowsplentyoftimeforlargelesslikelywavesfromthatseastateto ariseandbecountedintheregularwaveoccurrencedata. Forthisreason,theoccurrencesoflargeheightandlongperiodwavesare potentiallylargerthanlimiteddurationseastateswillproduce,becausealong seastatedurationisconsideredduringtheconversionprocess.Theselargeless likelywavesmaythencauseanincreaseinthefatiguedamagepredictedbyregular waveanalysis.
UGM2008FatigueReport.doc 6of10 18December2008

UGMSeptDec2008

Orcina
www.orcina.com

4.2. RegularWaveResults
Thefirstattemptmadeatregularwavefatigueforthissystemusedregularwaves rangingfrom1sto20sinperiod,combinedwith1mto20minheight.Overthis rangethediscretisationused1mstepsinheightand1sstepsinperiod. Therearetherefore400binsavailabletothewavescatterconversionmodulebut manyofthemareempty.The20m,20srangeswereusedbecausethemajorityof fatiguewavesshouldfallwithintheselimits. The1m,1sspacingfortheinitialstudywaschosensimplybecausethenumbers involvedwereeasytoworkwith.Asbeforewithirregularwaveanalysis,thisfirst passwillprovideaninitialdamagevalue,informationoncriticalloadcasesanda suggestedefficientdiscretisationforaccurateresults. Theregularwaveoccurrencetablegeneratedisincludedhereintheappendices.To representthe30stormscattercases,311regularwaveloadcaseswererequired. Thescattertablehastheexpectedoverallshape,witharoughlybellshaped distributionofwavecycles.Thevalueoftotalprobabilitycoverediscloseto1.0, showingthatthemajorityofwavesbelievedtoappearoverourchosen20year exposurearerepresentedinthetable. Thelifetimeestimatefromour1m,1sregularwaveanalysiswas182years, somewhatlowerthantheaccuratevalueof236years.Whenwelookattheother informationavailablefromthisanalysis,wecanformanopinionregardingthe reliabilityofthisresult. InasimilarmannertothedamagescattertableproducedforTable31,wecanview thecontributionofeachregularwavecasetothetotaldamagesufferedbythe system.Fortheregularwavecases,ratherthanshowascattertable,wesumthe damageoverwaveheight,andshowthevariationofdamagewithwaveperiod.
40% %-age of Total Damage 35% 30% 25% 20% 15% 10% 5% 0% 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 Wave Period Bins

Figure41Damagecontributionsfromregularwavesshowdependenceonwaveperiod UGM2008FatigueReport.doc 7of10 18December2008

UGMSeptDec2008

Orcina
www.orcina.com

Itisclearthatmostofthedamageiscontributedbyloadcaseshavingawaveperiod of9s.Thereisinfactaperiodrangefrom7sto12swherethemajorityofdamage isgenerated.Asmentionedpreviously,usingnumeroussmallwavebinswillyielda moreaccuraterepresentationofrealconditions.Whenresolvingasimplegraphof results,wetypicallyconsiderthat10pointswillprovideareasonablelookingcurve. Aswehaveacriticalperiodrangeof5swidth(from7sto12s)ourminimum resolutionoverthisrangewouldusuallyinclude10bins,eachof0.5swidth. Atthispoint,thefatigueanalysiscouldberepeated,dependingontheresultsas mentionedattheendofSection3.1.Inourcase,thisregularwaveanalysisdoesnot giveanacceptablefatiguelifeestimate,andsowerefinethebinsusedtodiscretise waveperiod,hopingforanswersclosertotheirregularwaveresult. Whereastheirregularwaveanalysisproposedthatthesecondattemptatfatigue analysiscouldreusepreviouscalculationsfornoncriticalcases,withregularwave analysis,thisislessimportant.Itwouldbepossibletodoso,buttheregularwave calculationsaresoquickthatitisprobablyeasiertorepeatthewholeanalysiswith afinerbindiscretisation,thantoperformaspecificsecondpassoverasmallperiod range,andthenaddbackinsomeoftheresultsfromthefirstattempt. Oursecondregularwaveanalysisused1mbindiscretisationforheight,and0.5s bindiscretisationforperiod,overthe20m,20srangesusedpreviously.Damage wasfoundtobelesssensitivetodiscretisationovertherangeofwaveheights,and soourapproachtherewasnotchanged.Therevisedfatiguelifeestimatefromthis secondregularwaveanalysiswas190years.Insearchofaconvergedregularwave result,afinalsetofcaseswasgeneratedusing1m,0.25sbindiscretisationoverthe 20m,20sranges.Thismostdetailedanalysisyieldedafatiguelifeof191years. Theconvergedregularwavefatiguelifeshowsascatterof19%awayfromourtrue value.Thisdistancefromtheacceptedtruevalueissimilartothescatterinresults obtainedbyrunning20minutedurationirregularstormcases.Assuch,regular wavefatigueanalysisprovidesanalternativetorunningshortstormdurationsfor thefirstloadcasescreeningfatigueanalysis. Finallywenotethattheconvergedregularwavefatiguelifeisshorterthanthe irregularwavepredictions,andthatthisconformstoourearliercomments.The conversionofirregularstormsintoregularwavecasescanleadtoamore conservativeanalysis.

5.

Conclusions

Irregularwaveanalysisisgenerallyacceptedasbeingamorerealistictreatment thanregularwaves,andassuchshouldincludelessconservatism.Thishasbeen confirmedbyourresults. However,wehaveshownthattheresultsofirregularwaveanalysisareaffectedby thedurationofthestormcasesused.Thedegreeofscatterthatappearsinrepeated irregularwaveanalysesindicatesthatastormdurationofonehourgivesaclose groupingoffatiguelifeestimates.

UGM2008FatigueReport.doc 8of10 18December2008

UGMSeptDec2008

Orcina
www.orcina.com

Becausefatigueanalysisiscomputationallydemanding,wewerelookingforwaysto optimisetheprocess.Forregularandirregularwavemethods,theconcentrationof damageoveraspecificwaveperiodrange,orinkeystormcases,wasvaluable information.Todiscoverwhichpartoftheinputisgeneratingthemajorityofthe damage,ascreeninganalysisisrecommended. Thescreeninganalysiscanbeperformedwithregularwaves;takingadvantageof theautomationfeatureswithinthescatterconversionmodule,orwithirregular stormsofshortduration.Bothapproachesareaccurateenoughtoprovideuseful initialinformationonthesensitivityofthesystemsfatiguelifetothevariousinput cases. Thesecondanalysiscanthenconcentrateonachievingaccurateresultswherethey willhavethemostinfluenceonthefinalfatiguelife.Thisanalysiswilltherefore usuallybeslowerthantheinitialscreening.Thepowerpointpresentationthat accompaniesthisreporthasmoreinformationonthefasterperformanceof OrcaFlex9.2comparedtoversion9.1. InOrcaFlex9.2,webelievethatthefatiguefacilitiesarenowfastenoughand sophisticatedenoughtoprovideanexcellenttoolforservicelifeassessment.We havetriedtoconsiderthewholeprocessofsettingupafatigueanalysis,inorderto provideadviceonhowtogetthebestoutoftheprogram.Ifyouhaveanycomments orquestionsregardingthefatiguemoduleinOrcaFlex,thenpleasedogetintouch. ColinLewis 18December2008

UGM2008FatigueReport.doc

9of10

18December2008

UGMSeptDec2008

Orcina
www.orcina.com

6.

Appendix:RegularWaveScatterTable

NumberofOccurrences Height (m) 20 19 18 17 16 15 14 13 12 11 10 9 8 7 6 5 4 3 2 1 Period (s) 1.7E01 6.0E+07 1 2.6E+00 6.2E+04 5.0E+07 2 9.1E01 1.5E+02 3.4E+04 1.7E+06 4.1E+07 3 3.4E01 3.0E+00 4.9E+01 1.1E+03 1.9E+04 3.1E+05 3.8E+06 2.6E+07 4 1.4E01 6.0E01 2.5E+00 1.1E+01 4.8E+01 2.4E+02 1.6E+03 1.4E+04 1.2E+05 7.5E+05 3.1E+06 1.1E+07 5 1.2E01 3.5E01 1.0E+00 3.0E+00 8.4E+00 2.4E+01 6.7E+01 2.0E+02 6.2E+02 2.1E+03 8.3E+03 3.9E+04 1.7E+05 6.4E+05 1.8E+06 5.1E+06 6 1.8E01 4.5E01 1.1E+00 2.6E+00 6.1E+00 1.4E+01 3.1E+01 6.7E+01 1.5E+02 3.2E+02 7.4E+02 1.8E+03 4.5E+03 1.3E+04 4.4E+04 1.3E+05 3.6E+05 8.9E+05 2.6E+06 7 5.8E01 1.2E+00 2.6E+00 5.3E+00 1.1E+01 2.1E+01 4.0E+01 7.8E+01 1.5E+02 2.9E+02 5.6E+02 1.1E+03 2.3E+03 4.6E+03 1.0E+04 2.7E+04 7.4E+04 1.8E+05 4.4E+05 1.4E+06 8 1.8E+00 3.3E+00 5.9E+00 1.0E+01 1.8E+01 3.1E+01 5.3E+01 9.0E+01 1.5E+02 2.7E+02 4.8E+02 8.9E+02 1.7E+03 3.2E+03 6.2E+03 1.3E+04 3.5E+04 8.9E+04 2.2E+05 8.6E+05 9 3.0E+00 4.9E+00 7.9E+00 1.3E+01 2.0E+01 3.1E+01 4.8E+01 7.4E+01 1.2E+02 1.9E+02 3.1E+02 5.3E+02 9.8E+02 1.8E+03 3.4E+03 6.7E+03 1.6E+04 4.5E+04 1.2E+05 5.5E+05 10 3.0E+00 4.7E+00 7.2E+00 1.1E+01 1.6E+01 2.4E+01 3.5E+01 5.1E+01 7.6E+01 1.2E+02 1.8E+02 3.0E+02 5.2E+02 9.7E+02 1.8E+03 3.5E+03 8.2E+03 2.4E+04 7.1E+04 3.7E+05 11 2.1E+00 3.2E+00 4.9E+00 7.3E+00 1.1E+01 1.6E+01 2.2E+01 3.2E+01 4.7E+01 6.9E+01 1.1E+02 1.7E+02 2.8E+02 5.1E+02 9.8E+02 1.9E+03 4.4E+03 1.3E+04 4.4E+04 2.6E+05 12 1.2E+00 1.8E+00 2.8E+00 4.2E+00 6.2E+00 9.1E+00 1.3E+01 1.9E+01 2.8E+01 4.0E+01 6.1E+01 9.6E+01 1.6E+02 2.8E+02 5.5E+02 1.1E+03 2.5E+03 7.9E+03 2.8E+04 1.9E+05 13 5.6E01 9.1E01 1.4E+00 2.2E+00 3.4E+00 5.1E+00 7.4E+00 1.1E+01 1.6E+01 2.3E+01 3.5E+01 5.5E+01 9.1E+01 1.6E+02 3.2E+02 6.5E+02 1.5E+03 4.9E+03 1.9E+04 1.4E+05 14 2.5E01 4.2E01 7.0E01 1.1E+00 1.8E+00 2.7E+00 4.1E+00 6.2E+00 9.2E+00 1.4E+01 2.1E+01 3.3E+01 5.5E+01 9.7E+01 1.9E+02 4.0E+02 9.4E+02 3.2E+03 1.3E+04 1.1E+05 15 1.1E01 1.9E01 3.3E01 5.6E01 9.2E01 1.5E+00 2.3E+00 3.6E+00 5.4E+00 8.3E+00 1.3E+01 2.0E+01 3.4E+01 6.0E+01 1.2E+02 2.6E+02 6.2E+02 2.1E+03 9.3E+03 8.6E+04 16 1.6E01 2.8E01 4.8E01 8.0E01 1.3E+00 2.1E+00 3.3E+00 5.1E+00 8.0E+00 1.3E+01 2.2E+01 3.9E+01 7.7E+01 1.7E+02 4.2E+02 1.5E+03 6.9E+03 6.9E+04 17 1.4E01 2.5E01 4.5E01 7.6E01 1.3E+00 2.0E+00 3.2E+00 5.2E+00 8.4E+00 1.4E+01 2.6E+01 5.2E+01 1.2E+02 3.0E+02 1.0E+03 5.1E+03 5.6E+04 18 1.4E01 2.5E01 4.5E01 7.7E01 1.3E+00 2.1E+00 3.4E+00 5.7E+00 9.7E+00 1.8E+01 3.6E+01 8.3E+01 2.1E+02 7.6E+02 3.9E+03 4.6E+04 19 1.5E01 2.7E01 4.8E01 8.3E01 1.4E+00 2.3E+00 3.9E+00 6.8E+00 1.3E+01 2.6E+01 6.0E+01 1.6E+02 5.6E+02 3.1E+03 3.8E+04 20

UGM2008FatigueReport.doc 10of10 18December2008