Effect of lateral impact loads on failure of pressurized pip(7)

Effectoflateralimpactloads199

thepipespecimen;DmaxandDminthemaximumandminimumdiametersofthedeformedpipecross-section;WbthetotaldisplacementofpipebottomsurfaceattheimpactlocationCfthemaximumper-manentdeformationoffoundationattheimpactpoint;andlAthetotalrelativedisplacementofpointsAatbothsidesofthesupportsasshowninFig.3(a).ThevaluesofwinTable2arethemoisturecontentofsoilfoundation.InTable3,lBisthetotalrelativedisplacementofthepipeslidingoutoftheclampsatboththesupports.

ThevaluesofthetotalmaximumpermanenttransverseplasticdeformationWfofthepipespeci-menslocatedattheimpactpointwererecordedinTables1to3.Thesevaluesweremeasuredaccord-ingtothesupportpositionacrossthespan2LbeforeandafterunclampingfromtherigasWfoandWf1,respectively.However,therewasnotmuchdifferencebetweenthesetwovalues.Speci-menBLP3showedthemaximumdeviationof7.18percent(i.e.2.50mm),whereasspecimenCHP5showedtheminimumdeviationof0.08percent(i.e.0.02mm).

Threefailuremodeswerecarefullyobservedinthepresentimpacttests,asidenti edinreference[6].FailuremodeIcorrespondstothefailurewiththeshearslidingorindentationattheimpactpoint.Thephenomenonofbucklingatthebottomsurfaceofthepipenearthesupportsisassociatedwithfail-uremodeII.FailuremodeIIIrelatestothetensiletearingonthetopsurfaceofthepipenearthesupports.

ThesethreefailuremodesareshowninFig.5andrecordedinTables1to3foreachgroupofthepipespecimens,wheresymbolIdenotesthatthepiperemainsintact(modesIandII);Jindicatesasmallcrackorjusttheonsetoffailureoccurringatthepipenearthesupports(allthreefailuremodes);andCrelatestothefailureofpipewithalargecrack,i.e.crack.3mm.Thecriticalinitialimpactenergyisde nedastheaveragevalueofthesmallest

energythatcausedpipefractureandthelargestenergythatdidnot.Noneofthepipeswasobservedfracturedunderneaththeindenter,asthecylindricalimpactsurfacewasused,insteadofa atimpactsurfaceatthetipoftheindenter.

TheresultsfromtheimpacttestsshowedthatonlyfailuremodeIIIcausedalossofthepipesintegrity,andthisoccurrednearthesupportsasreportedinreference[6].WiththeexceptionofspecimensA1,A2,A4,andBHP3,itwasfoundinthepresentexper-imentalstudythatfailuremodeIIImostlyonlyoccurredononesideofthepipespecimens.Besidesthis,thisfailuremodejustoccurredontheinletsideofthenitrogengasonthepressurizedpipespeci-mens.FailuremodesIandIIoccurredinalltheimpactedpipespecimens.

Onacarefulinspectionofthedeformedpipespecimens,itwasfoundthatthemiddleaxiallineofthepipespecimenwascreatedbynearlytwostraightlinesthatconnectedattheimpactpoint,asreportedinreference[6].Thisphenomenonwasusedinthetheoreticalanalysesofreferences[4,8].

33.1

FINITEELEMENTFAILUREPREDICTIONFEAmodelling

Fig.5

Threemajorfailuremodesoccurringatthepipe:(a)shearslidingattheimpactpoint(modeI)and(b)bucklingonthebottomsurface(modeII)andtensiletearingonthetopsurface(modeIII)

LUSASV13.4isageneralFEApackage,whichcannotdirectlyestimatethefailureofthepipelinesunderimpactloading.Therefore,FEAcalculationswerecarriedoutforthemostpipespecimensusedintheimpacttestsinordertocomparewiththeexper-imentalresultsforassessingtheaccuracyofthenumericalanalysis.Arangeofpracticalimportantconditions,includinginternalpressure,foundationsupport,andpipedimensions,weresimulatedasintheexperimentsinsection2.Furthermore,boththegeometricalandthematerialnon-linearitieswereconsideredwiththeintentionofsimulatingtheactualdeformationpatternand ndingoutthestressandstraindistributionsintheareasofinterest,wherematerialruptureoccurredbecauseofthehighstrainconcentrationinthepotentialfailureareas.Asaresultofthesymmetricalnature,onlyone-quarterofthestructureswasexamined.ThepipeswithH¼1.6mmshouldbedividedintofourlayersalongthewallthickness[7].ThelengthoftheruptureareanearthesupportandthecontactareasbetweentheindenterandthepipeinthepipelongitudinaldirectionshouldnotbelessthanthepipemeanradiusR[7].

Non-linearelasto-plasticisotropicmaterialprop-ertieswithstrainhardeningwereusedforallthemildsteelpipesinthepresentsimulations.Thesewereconstructedaccordingtothevon

Mises

Proc.IMechEVol.220PartE:J.ProcessMechanicalEngineering

JPME97#IMechE2006