Application-of-CFD-(Fluent)-to-LNG-spills-into-geo

Availableonlineat(2008)158–168ApplicationofCFD(Fluent)toLNGspillsintogeometricallycomplexenvironmentsFilippoGavellia,∗,EdwardBullisterb,HarriKytomaaaaExponent,Inc.,17000ScienceDrive,Suite200,Bowie,MD20715,UnitedStatesbCambridgeTechnologyDevelopment,Inc.,UnitedStatesReceived4February2008;accepted5February2008Availableonline17February2008AbstractRecentdiscussionsonthefateofLNGspillsintoimpoundmentshavesuggestedthatthecommonlyusedcombinationofSOURCE5andDEGADIStopredicttheflammablevapordispersiondistancesisnotaccurate,asitdoesnotaccountforvaporentrainmentbywind.SOURCE5assumesthevaporlayertogrowupwarduniformlyintheformofaquiescentsaturatedgascloudthatultimatelyspillsoverimpoundmentwalls.TherateofspillageisthenusedasthesourcetermforDEGADIS.Amorerigorousapproachtopredicttheflammablevapordispersiondistanceistouseacomputationalfluiddynamics(CFD)model.CFDcodescantakeintoaccountthephysicalphenomenathatgovernthefateofLNGspillsintoimpoundments,suchasthemixingbetweenairandtheevaporatedgas.BeforeaCFDcodecanbeproposedasanalternatemethodforthepredictionofflammablevaporclouddistances,ithastobevalidatedwithproperexperimentaldata.ThispaperdescribestheuseofFluent,awidely-usedcommercialCFDcode,tosimulateoneofthetestsinthe“Falcon”seriesofLNGspilltests.The“Falcon”testserieswastheonlyseriesthatspecificallyaddressedtheeffectsofimpoundmentwallsandconstructionobstructionsonthebehavioranddispersionofthevaporcloud.Mostothertests,suchastheCoyoteandtheBurroseries,involvedspillsontowaterandrelativelyflatground.ThepaperdiscussesthecriticalparametersnecessaryforaCFDmodeltoaccuratelypredictthebehaviorofacryogenicspillinageometricallycomplexdomain,andpresentscomparisonsbetweenthegasconcentrationsmeasuredduringtheFalcon-1testandthosepredictedusingFluent.Finally,thepaperdiscussestheeffectvaporbarriershaveincontainingpartofthespilltherebyshorteningtheignitablevaporcloudandthereforetherequiredhazardarea.ThisissuewasaddressedbycomparingtheFalcon-1simulation(spillintotheimpoundment)withthesimulationofanidenticalspillwithoutanyimpoundmentwalls,orobstacleswithintheimpoundmentarea.©2008ElsevierB.V.Allrightsreserved.Keywords:LNG;LNGhazard;Spills;Vapordispersion;CFD;Impoundment1.BackgroundandmotivationNaturalgasdemandhasbeenrapidlygrowingworldwideoverthepastfewyears,leadingtoalargenumberofapplicationsforthesiting,constructionandoperationofnewLNGreceivingterminals,particularlyinEuropeandNorthAmerica.Currentlythereareapproximately40newLNGreceivingterminalseitherinconstructionorinthepermittingprocessinNorthAmericaalone,andapproximately20additionalpotentialsitesarebeingconsidered.ThegrowinginterestinLNGandthepotentialforsomeofthenewterminalstobelocatedinproximityofhighlypopulatedareashasraisedquestionsaboutthesafetyofLNGtransportationandregasification.∗Correspondingauthor.Tel.:[phone];fax:[phone].E-mailaddress:[e-mail](F.Gavelli).AnimportantelementoftheoverallriskassessmentofLNGoperationsisthedefinitionofthehazardfootprint—thearea,centeredaroundtheLNGterminal,withinwhichhazardousconditionsmaybepresentintheeventofanLNGspill.Thehazardousconditionsofinterestforpublicsafetyare:(1)radi-antheatfluxfromanLNGpoolfire;and(2)thepotentialforaflammablevaporcloudfromanLNGspilltoreacharemoteigni-tionsourceandresultinaflashfire.Federalregulations(49CFR193)requirethatapplicationsforland-based(onshore)LNGreceivingterminalsdemonstratethatthehazardscreatedbyanLNGspillwillnotextendbeyondtheareaunderthecontroloftheterminaloperator.Inthispaper,weaddressthedispersionofaflammablevaporcloudfollowingaspillofLNGintoanimpoundment.Todate,themodelingtoolsusedtoestimatethedispersionofvaporsemanat-ingfromanLNGspillhavemostlybeenlimitedtointegral-type0304-3894/$–seefrontmatter©2008ElsevierB.V.Allrightsreserved.doi:10.1016/j.jhazmat.2008.02.037F.Gavellietal./JournalofHazardousMaterials159(2008)158–168159models(e.g.,DEGADIS[1]),whoseassumptionslimittheirusetodispersionoverflatterrainorwater.Typically,thecalculationofflammablevapordispersionhazarddistancesforLNGspillsintoimpoundmentsutilizesDEGADIStogetherwithanotherintegral-typemodel(SOURCE5).GiventheLNGspillflowrateandthesizeandmaterialsoftheimpoundment,SOURCE5cal-culatesthetime-dependentvolumeoccupiedbythenaturalgas(thesumoftheresidualLNGvolumeandthevolumeofgasproducedbyvaporizationoftheLNG),andcomparesittothevolumeoftheimpoundment.WhenthecombinedvolumeofLNGandsaturatedvaporintheimpoundmentexceedsthevol-umeoftheimpoundment,theexcessvaporisconsideredtobereleasedoutoftheimpoundmentandbecomesasourcetermforDEGADIS,whichcalculatesitsdispersion.RecentdiscussionsintheLNGsafetydebatehavepointedoutsomelimitationsintheSOURCE5model,namelyitsinabilitytoaccountfortheexpansionofthevaporvolumeduetoheatingabovetheboilingtemperatureofLNG,aswellasitsinabilitytoaccountformixingbetweentherisingvaporcloudwithintheimpoundmentandthewindovertheimpoundmentwalls.Con-sequently,theFederalEnergyRegulatoryCommission(FERC)hastakentheconservativepositionthattheuseofSOURCE5toquantifytheLNGvaporsourcetermforuseinDEGAD