BRITTLE FAILURE IN CARBON PLATE STEEL …

1 , zABSTRACT _'~hefailureofshipsat,sea~~ll(iatdocksid etlurirlgWc]rl(lW~l,],IIt]roLlght,thepro l] ~fromshipfailureshavebeenwellcorrelated, andasaresult,muchhasheel] , ,plusfuiluresingastransmissionlines, ,pressurevessels,asmo kestack,apmstock,powershovels, :(1)Brittlefailureillnonshipstructuresis thesamephenomcnonasoccursinships;(2)brit tlef~ilureoccursinmanytypesoflmnshipstru ctures;(3)brittlefracturescancrossrivete djoints;(4)thereisnoevidencetoshowthatth epercentageincidenceofbrittlefailurehase itherdecreawdorincreasedwiththeadventofw elding;(5)inconjunctionwithotherfactors, thermidstressmaybcimportant;(6)residuals tressesm-enottheprimemuseofbrittlefailur e,butsuchstressesmtiy,inconjunctionwitho therfactors,i~litiatesurhfailure;(7)thee ffectofmetallurgicalvariablesisimportant ;(S)coldformingpromot,essusocptibilifiyt obrittlefailure,butitsrolecannotbeassess edductolackofdata;(9)insuchcaseswheredat aareavailable,Chm-plyimpactvtiluesofplak :weregenerallylowtitthefailuretemperatur e;(10)inmostrases01nonshipbrittlefailure ,thefractureoriginatedatdefectsarisingfr omftibrication,Afew-originwkdatdesigndef ects;(11)itseemscwidentinallcasesthatfra ctureoriginatedata~cometriccliscont,inu- ity;(12)nocvidenccexis(,sforthesefaileds tructurestoshowtheeffeetsofvariousweldin gprocessesonsusc~ptibilitytobrittlef~~il ure;(13)exceptinthecaseofex-ceptionallyp oorwelds,thereisnotendencyforfracturet,o foilowweldedseams;(14)thegreatmajorityof nonshipbrittlefailuresapparentlyorcurund erconditionsofentirelystaticlofiding;( )agcofstructureseemstohavenobearin~o~lbr ittlefailure;(10)mostengineeringcodesper mittheuseofsteelwhichisknowutobeparticul arlysusceptibletobr~ ,underallcodesbutone,thestresslevclfiare heldtoq~liteconservativevahles.)))

2 (17)finally, ~wrenee,Lndthereisnoknowntest,whichwills urelypredictfromthebehaviorofsmfillspeci mensthepcrforrnanceofagivensteelincircum stanceswheretitructurnibrittlefailuremig htoccur,111short,carefuldesign,selection ofmtiterials,andK(IOCI workmanshipareofthegrmtcstimportanceinth epreventionofbrit~ :lctiol~.. ~..56 WaterSta~~dpipes, ~ ~ ;..j,.. ~]IIlstorlesofErlttlcFwllurcs rhcEraofWeldedConstruct~o~ .- ..-17 Pressure> ~ ~ ~~ ~ .- .. ~[ - - - :; ~ckgroundofI~ ~ ~et, {:tu~.FZUlures~,,..34C{]mparlsol~[Jflall ureInc~de~ +;ffect, - -~(j,..Effectotl%esldu~ ~..Effect,ofMet,allurg]calV~ir~ablesandC hcmlcalComposltlo] ~Cold~ ~ :.:..40 TheR,oleoiCracks,StressC~oncentratlons,a nd(ltherDefectslnI] ,,,..,,.,,,.,,..,~.~.:.. ~ ,. 46A$~ ~ ( )UCTION1 NYcriticalsumwyofbrittlefailuresofcarbon steelplate,nonshipstructurefimustnecess~ wilybeori- ailllreproblemwasbroughtintosharpfocusdu ril~gWorldWarIIwithth(~breakingupatseaan datdocksideofweldedmerchantvessels, ]t,muchlighth~ ormatio~lrelatingtofailuresexistsintheca seofnonshipst,ruct~~ ~l,therandcorre-latesuchdata, :a1]onshipsurveyrevealsho])-\\-idc-sprca disthebrittlefailureproblcrnillnonshipin dus-tries,howlo~lgtheproblcmhascxistmlj: mcitowh~ ,ishoped,moreover,thatthepublicationofsu chflSLIPV eywillhelptocategorizeandse~forththecirc umst:allcesinwhichbrittlefailuretiarclik c]ytoorcur,Itmightbewell,themfom, ~gshout, ,1 OJV-Lrmpcraiure,suchasexist,sinthearnbic ntat, anotch(introducingtriaxialstress).)))}}

3 Anydefert,,s~lchasaweldingcrack,orvoid,o racrackleftbyaplll~chil~gorshearingopera tion,ranserveasanotchwhichwi]1initiatebr ittlefailure,Thllsllrit,t]efailllreissom etimescalled notchbrittleness, Thethirclf art,orishighstrainratleorimpactlo:lding, Thisthirdfactor,how- ____ __evw,i,snot, ,manybrittlefailureshavebee~linitiatedun derwhat, ~~ilureoccursii; (approwjhillgseveralthousaudIcetpersemnd ),almost,completelackofductility,negligi bleencrgvabsorption,andabrittleorfareted fippeartinccofth~fracturedsm-fac~.Monwwe r,thefract~u-edsurf:i,ceoft,cnhasachm-:w tcrist,ic [chevron or hcrringbont+ appearanccjtheapicesoftheherringi]onospo intiugtotheoriginoithefractl~ ,ifcxi,lrcrneexampleof itwatithislwtcharacteristicthatwassobaff iingtothee~~gin(~erswhofirst, (jjtheT3essemcrprocesfiofsteclm:~kingwas announcedtothew-or]d,andshortlythereatte rst,eclbemmeavailableiucomparativelylarg equantities,Afewyearslater(1S61)theopen- hearthprocessb( (blisterbtir).Ttwmsrarceaudexpeusive, ,whichbecauseof itsslfigil~cl[[sionsisMlextremelytoughma ,-terialjwasusedfolstructurt~[ ,13esscmersbcelwastivailal>lcinsurhq(lau tit,vthatitw-w+usedforboilerplate,:,wdin 1863 G4twoslmelvesselsof377tonsa,lld1283tolls werebuilt01st( ,aslateati1877,Boardol TradeRcgulatio~ltiprohibitedtheuseofstee lincomstruct,ionjfindremovaloftheseregl~ la[ionsinthatyearprovidm]LLgreatstim~llu stothesteelilld~ititryinthatcoun~ry.)))) )

4 1 h{wdllriugtheperiodof1860to1890, ,cdSt,atcs,~vroughtironwasgradu:dlyheill gsupp]alltedasaStrurt,llr:lltnct,a] !-hepartofellgincorstodificardareliablem :~,teri:,dlike]vroughtironcauseclthechaI lgetoromeabo~ ,thecheuplmss,greateravailability, ,otN&ACADEMYOFSCIENCCS-NATIONALRKs~RcH =o~NclLZIO1 CONSTITUTIONAVENUE,WASHINGTON25, .. ,trouhlcs~ ,presentedafiameeting thatyem,byNathanielJ3ar nabyon TheIJseofSteelinlNavalConstructioti. I+ ] ,inthesamediscussion, ,hecitesasteelpltitethat whencold,onbeingt,hrowmdowrl, , ~ouldsplitwith~ ,Mudifaremedywasnota,vailablc,atleastara tional~xplana-t, , @oreexaminingindetailthehistoryofmmshipb rittlefailures,itiswelltoglanceforamomen tatthesta,tistics $ ,942-52about250weldedshipssufferedoneorm orebrit,tlefracturesofsuchse-veritythatt hevesselswerelostorinadangerouscondi-tio n,Nineteenofthese250shipsbrokecompletely intwo, , ~thesame10-yrperiod, ,germrally1essthan10 ,but, ,however, ,hetankertype,thesamecategorywhichgivefl themost, ~ ,indicatingbrit, ,frequentmentioninthetechnicalliterature ofcracksinnumer-ousrivetedvesselsindicat estheprevalenceot ,thebrittlefailureofcarbonsteelplate,non shipstrwctures, carbonsteelplate , , ] , , ,numerouslettersweresenttovariousindus-t ries,tcxhnicalorganizationsandgovernrne~ (PIUS probablegaslinefailures) (inadditiontofailuresongaslines) ~ present, , ,forvariousreasons, ,whentheyoccurred,Wer(?)}

5 Probablyatiributcdtoothercauses,suchasfa tigueor,aswillIatwbeshown, badsteel. Moreover,mostindustriesinthepastwerenota nxioustorevealacci-dentsofthistype, ,unlesspersonalinjuryorpropertydamageres ulted, ,eutirelybygovernmentagencies, , ,however,thattwoofthelargest, ,theseparticularindustriesareinmoredires traitsthananyother, THEERAOFPREWELDEDCONSTRUCTIONS ignificantbrittlefailuresofsteelplat,est ructuresinthepreweldedperiodprovideausef ulbackgroundtopresent, >fracturedifficultiesdidnotbeginwitht, ,Gruvesend,LongIslqnd, , , , ,decreasingconicallyinalengthof25fttoan% ,5by7ftand5by9ftwereemployed, , states: Someplatesarebentalmostdouble,andothersa reactuallyrolledup,showingaverytoughmeta l,..Theutterde-structionofthelowerparts0 1thetowerandtheappear-anceofthefallentow er,whichisbrokenintw-ojustabovetheconean dpresentsanalmostcleansquarecutjustbelow thiscone, ~m-minguponthegeneralevidence,weshouldsa ythattheplateswereamplythickenoughtostan dthestressputuponthem,evenweretheyagoodw roughtiron;theworkmanshipseemstohavebemg enerallygood, , ~ ebroughttheutterdestructionthereexhibite danditwouldseemasifthisbrittlematerialha dunfortunatelybeenconcentratedintheporti onofthetowerexposedtothegreateststrain.

6 Thepresent-dayengineerimmediatelynotesth atsomeoftheplateswereveryductile,othersa ppearednottobeductile,andthatthereporter onthescenebelievedthatmanybrittle(andthu sdefective) ,Brooklyn, , ,1898 , , : .. ~structureatthecenteroftheplate,shadingo ffintoaveryfinegrainatthesurface,withher eandtheresplinteredgesmuchlikeabrokencas e-hardenedmaterial. zThewitnessurgedasearchinginvestiga-tion intospecificationsfortheplates,theirchem icalcomposition,andbehaviorundertest( ,tensiletest). ,Sanford,Me., , ,19043 Thiswasarivetedsteeltank,40ftindiameter, ~thicknessvariedfrom , ,anditW-asnotedthatmanysma] : ..enoughcleanfractureswerefoundtoindicat ethatthesteelwashardandbrit-tie, ; ; [holes].Itisnotevident,however!whyfailur edidnottakeplaceim-mediatelyupontheiniti alapplicationoffullpressure. 3 Ormparagraphoftheaccountdescribesinexcel lentdetailwhatisnowcalledthe shearlip. ,Boston,Mass.,12 Noon, ,19194 9 TheBostonMolasses Tankexcitedgreatinterestatthetimeofitsco llapse, 16ontheBostonwaterfront, ,50fthigh,withlap-jointedplatm / , ,300,000galofmolasses, , , , ,bothonthefactsofthecaseastheysawthem, ,6showedthat,atthebaseofthetankwhenfull, stressinthethickerplateswas26,400psi,and inthethinnerplates26, %,stressesinthejointwere40,000to50, ,withbombssubmergedinmolasses, ,andotherstestifiedthat,thetankwasstruct uralysound,thatitdidnotruptureatitsweak- estpoint,thattestsshowedthematoria,ltoha voaten-silestrengthof55,000 56,000psi,andt,hatallp] ( ) ,ofHarvard, (crystallographictwinning)foundinthemicr ostructurewereusual] ,particular] ,obtainedinlaboratorytests,showingshort, sharp,hcrringt)on~dfractures, ,ofMITtcstifiwithatNeumannbandswouldappe arintheta~lkfracturesduesolelytotheactio nofmolasses.

7 *Otherstestifiedthatthetankdesignwasunso und,thatpunchingofrivetholeshadstartedsh ortcracksandthattheta, ,afteryearsoftestimony,thecourt-appointe dauditor, (lgden,handeddownadecisiongthatthetankfa iledbyoverstress, ssummaryisworthreproducinghere,sinceitfa irlywellsummarizestheknowledge(orlackofi t),thencurrentamongpracticingengineers,c onmrningnotchbrittlebehavior..Thedefend- ant sexpertscalledattcmtiontothepresenceofNe u-mannbandsinsteelofthecharacterhereinco nsideredwhichhadbeenfracturedwasaproofth at, .. ,,.. ,,., indopendenfidiscoveryofthemeanin~ofchevr onmarkingsisoftheearliestonrecord! ,however, ,88inlob-omtorytes~iugof,numcrousrmiteri rdsnotedchevronmnrkin~s,andthefnctthatth e& ,however, ,ofthestrengthofmate-rials,oftheforceofh ighexplosives, [ddbeandinthiscasewereproducedbythepress urecausedbytheweightofthemo] ~ , ,swearthatthesecondarystressesinastructu reofthiskindwerenegligibletindIhaveheard fromequallyauthoritativesourcesthat, , ,however, ,PoncaCity,Oklu., , !J,19z5 MThistankwas117ftindiameter, , , Fthedaybeforethefailure,andhadsuddenlydr oppedto 4 ,muffledsoundswereheard, ,anddecidedthat, , ,however, (XlStorageTanks, , ,Early1930 s ~ , ,of55, , , , ( )asfollows:NOTE:Fiquresshowmaredistances measureddownFrombori%on+alline+0+opOFbo+ +om5 pla+ + ($,bestrecord10-+tJ15 + \i+ belo :%i+ch%naingTake.)))

8 # 997% \I\9\a-10)7\ ~.2F i,Soundin~Taken8% L-~ ~ , +breUndermadyo! Below,, ~ - s-z ~l4-20 *.3-~2-a,1e%<%undin$lakan~/8% 911/101 /\Lowg,~ ; ,1929r %mdin$Taken2-belowon19*-I#abnveonIWNew+a nkbottmnandbo++ +ionundernotih~ideof+aok. * Oooobooooomo000000+000020000 w00l&AL-llOOO O08000000* < ~l0oemmTaken ,1929 2 , ,1933 10 ,afterwhichanaproncoveringtheanglewaswel dedtoth~ ,3by3by5 , anlcNo.~,riveted,wasalso0155,000bblctipa city, ,failedat,9 ,, ,1933,ata,tempmatureof 4 , ~allbuttwelded,andasecondsetofangleshoes wasinstalledovertheoldshoesontopofthenew bottom.( ). ,theverticallegsinbothoftheseshoessplitd i-rect] ~bottomangleironwasbrokenatthispoint, ,where weldedtotheangleiron, ,thoughbynomeanscertain, ~.lbsewereerectedin1923, ,thelowercourseplatewas5 , ,withl/2-in, strapsincludedonlytworowsofrivets, , ~ ,rivetsbetweenthefirstandsecondcourses, ,1932,whenthetemperaturedroppedtoabout 18 , ,theverticalcrackpre-sumablystoppedmomen tarilyandthecrackedsheetsbegantoleanoutw ard, ~ I:ShunCROSSSECTIONOF130 TTOMANGLES&SHOES( )Fig.

9 ,anddetailoJ+failu, , , , ,andthetemperaturehadfallentobelow 30 1?. , , + ,17Mn0, , ( ,valueswere3to11ft-lbat, F,5ft-lbat0 Fand1to2ft-lbat 25and . ,120ftindiameter, , , ,1937, ,withtemperaturesrangingfrom 2!! * ,ordidnotoccur, ,whichoccurredinthe1stcourse, ,andinthifirespectit, wasalso120f tindiameterby40fthigh, ,througharivethole, , ~ !3%1showedcracksin3lowercoursesheets, ,afailureoccurredinoneofthesepreviouslyp atchedsheets,about6ftfrom ,120ftindiameterby40ft, ,1934, +5to 20 ,inpart, ,particularlysinceit,wasfoundthatthe\ ]] ,crudeoilisknowntopar-tiallysolidifyagai nsttheinsidewallsandbottom,thusinsulatin gthetankwallsfromthewarrrwoilatthecenter andenablingthetankshelltoapproach,prob-a blyfairlyclosely, ,however,beingprotectedontopandincontact withrelative]ywarmgroundunderneath, ,therefore,atemperaturedifIerenccbetween thecenterandperipheryofatankbottomcouldr eadilybeoftheorderof50 ,namely, ..that, ,thejointbetweenthebottomandshellof:~tan kisoncofthemostcritical(ifnotthemostcrit ical) ,it,wouldseemthattheutmostcareshouldbeus edinthefabricationanc1inspectio~~ofthisj ointtomukemrethat,ibisassoundandfreeofde fectsmdotherstressraisersaspossible.)

10 ,MiddleWest, , , ,19@~zG!Ilisoiltankwasofrivetedconstruct ion}114t tindiMnetw,and30{ hcroofwtisofwood, ~.inthidmess,withwsing]erowofrivetsinthe hori-zoll~alsc~~ )} erequadrup] ]atesweresuccessivelythinner,thetop(fift h)coursebeing1/tin.~ ,alargetriangularhol~wa~CU~illth~~O~tOmc ourwofplateswith:1~ ~ornwasrepaired,thepiermofsteelwhichhadb eencutoutwaselectricallyweldedbackintopl ace. ,,<,,! J?, , ~teclearlyshowsinthefirstr{ ~ch, , , , ,andthebottomapexofthetrim~ ) ,theweldingofapatchintoasolidplateisknow ntoresultinahighcle-greeofconstraint, ,combinedwiththelowambienttemperature,wa ssufficienttoinitiateabrittlefailureinth eweld( )whichthenpropa-gt~tedup~r~ddownt, ~NE~/~ : :T;;TLi%?GEDATEOFUETRESREMARKFEETIHASSEL T~ E- y%%E:6 ::: %! _y4 &,, -LL?RIOGEOEck2 E%WL -k 0:$$$s ;395 936,37%?r;16-73 11 \6-75n.!,,-- ----+-54:,,.*wEiOEO200!23 KAULKLEF@=== ~mM34,3S ::;sJ99L--::9g;; (l?o{luwGmz,Ii,72,94JU,,Y14,1340)Fi, [hreeJailedVierendecl[.r,[ , AlbertCanal,Has-selt,Belgirl,m, , ~93fi~2 04 ThisbridgewasofatypeknownasaVierendecltr uss, ,lowerrhords,withcurvedupperchords, 1 heupperandlowerchordswerejoinedbyvmtical s.}