Evaluation of Fillet Weld Requirements (1501-S)

1 Project Summary Report 1501-S 1 The University of Texas at AustinCenter forTransportation Research PROJECT SUMMARY REPORTCENTER FOR TRANSPORTATION RESEARCHTHE UNIVERSITY OF TEXAS AT AUSTINP roject Summary Report 1501-SFillet Welding Procedure Qualification ResearchAuthors: Heather E. Gilmer and Karl H. FrankCenter for Transportation Research, The University of Texas at AustinDecember 2001 Evaluation of Fillet weld RequirementsThe motivation for thisresearch was the desire todevelop improved procedures forqualifying Fillet welds on bridgestructures. The current procedurequalification tests prevent the useof active fluxes and otherconsumables or procedures whichmay be more applicable to filletwelding. Active fluxes areformulated for limited-passwelding. They contain activedeoxidizers, such as manganese,silicon, or both, to improve theresistance to porosity and weldcracking caused by contaminantson or in the base metal.

2 Mostfillet welds are single-pass weldsapplied to unprepared enhanced ability of activefluxes to deoxidize the weldmetal is particularly important forfillet welds. The amount ofmanganese and silicon in theweld metal varies with the arcvoltage, and so the arc voltagemust be carefully controlledwhen making multipass weldswith active fluxes. The change inthe amount of silicon andmanganese when the arc voltageis changed is used as an index todifferentiate between active andinactive or neutral fluxes. Moreactive fluxes will show a largerchange in deposited weld metalchemistry for an incrementalchange in Fillet weld qualificationrequirements in the current bridgewelding code, ANSI/AASHTO/AWS , henceforth , specify that Fillet weldingprocedures be qualified using agroove weld specimen ( , Section ). A largegroove weld is used to producethe test specimens. The weld isdesigned to provide as near aspossible a weld that is undilutedby the base metal.

3 Fillet weldsare often single pass welds thatcontain a considerable amount ofbase metal in the cast welds. Atypical small Fillet weld will havemore dilution of weld metal withbase metal than the material atthe center of the large grooveweld used in the standard test. Inaddition to the difference in theamount of dilution, the grooveweld microstructure will berefined in subsequent passes;single-pass Fillet welds undergono refinement. In practice,welding procedures that givegood test results for a grooveweld may not necessarily producethe best Fillet welds. In particu-lar, fabricators have reported thatthe heat input required to producea groove weld specimen that willpass the specified tests is too highfor many Fillet welds. Thisrequirement is particularlyproblematic with T-joints weldedsimultaneously on both sides,where the total heat input to thewelded area is greatly are anecdotal reports thatfillet welds made with proceduresthat pass the qualification testshave failed in the research investigated thebehavior of Fillet welds todetermine what Requirements ifany should be imposed upon thefabricator to ensure the satisfac-tory performance of Fillet weldsused in bridges.

4 The research wasrestricted to welds made using thesubmerged-arc We research examined theperformance of Fillet welds madewith a matrix of consumables andheat inputs. Different fabricatorsmade the weld specimens usingconsumables they normally use intheir shop. All of the consum-ables were from the LincolnElectric Company. The weldmatrix is shown in Table 1. Thefabricator that provided theweathering specimens uses thisset of consumables for all hissubmerged-arc welds. This set ofconsumables was included sincethe fabricator does the majorityof the steel bridges in Texas. Thetwo heat input ranges used foreach set of consumables boundthe values that would be used innormal fabrication. The high heatinput values used for the activeflux were much higher than thefabricator would use in normalpractice. Two-sided as well assingle-sided welds were includedin the fabrication of the T-bendand the Fillet weld shear speci-mens. These welds simulate thewelding of a stiffener to web.

5 The3/8 and 1/2 in. thick platesProject Summary Report 1501-S 2 forming the stem of the T, thesimulated stiffener, were used todetermine the influence on theadditional heat input from the weld onthe opposite side upon the fabricator produced threespecialized Fillet weld test transverse shear specimen similar tothe specimen in AWS wasused to measure the shear strength ofthe weld . A T-bend specimen, whichhas been used by the Georgia andCalifornia departments of transporta-tion and also used to evaluate filletswelds used on the new high perfor-mance 70 grade bridge steel, was usedto measure the ductility of the weld root Charpy V-notch,WRCVN, specimen was utilized tomeasure the notch toughness of asimulated Fillet weld . In addition, astandard AWS groove weld qualifica-tion specimen was made usingweathering consumables at the highand low heat input in order to com-pare the results of the specializedfillet weld tests with the standardAWS specimen.

6 Three replicatespecimens were tested for eachcondition. The factorial experimentalresults were analyzed using analysisof variance We results of the tests indicatedthat the strength and ductility mea-sured in the shear and T-bend speci-mens were similar. The T-bendspecimen did not provide meaningfultest results. However, it did provide ameans of assessing the depth ofpenetration of the weld heat input double-sided dartwelds with an active flux producedcomplete penetration with a 3/8 cracking across the weldoccurred in these specimens. How-ever, the ductility of the single-sidedhigh heat input weld was the leastcracked. The hardness of the higherheat input welds was less than thelower heat input and the double-sidedwelds produced the lowest variation of hardness with heatinput was largest in the welds usingan active flux. The double-sided highheat input Fillet weld on a 3/8 in. stemproduced the lowest weld hardnesswith active and neutral fluxes.

7 Thespecimens using the weathering fluxshowed smaller variation in weldhardness with changes in heat estimated tensile strength for theweld with lowest hardness is 84 ksi,far above the required strength of results of the weld shearstrength tests showed for all consum-ables that the low-heat welds arestronger and harder than high-heatwelds and single-sided welds arestronger and harder than dart the calculated shear strength andthe tensile strength corresponding tothe hardness are well above thenominal tensile strength of 70 ksi forall speci-mens tested. The measuredshear strengths were as large as two tofour times the nominal value of x70 = 42 all three sets of consumables,no effect of heat input was foundwithin the dart-welded finding may have to do with theeffect of dart welding on actual heatinput. It is possible that althoughraising the heat input may changeweld strength, once a saturation heat input is reached there will be nomore effect from further heat inputincreases.

8 If this is so, then dartwelding will have no additional effecton a weld whose heat input is active flux specimens gavethe lowest absorbed energy in theWRCVN tests. A comparison of thelow heat input results is shown inFigure WRCVN, PQR, and certifi-cate Charpy that the consumablemanufacturer reported results areessentially the same for the active andneutral fluxes. Only the weatheringconsumables show a significantdifference between the WRCVN specimens and the groove weld testplate specimens performed for PQRand certification testing. The heatinput had little effect on the resultsfrom either the WRCVN or thenormal CVN test weathering consumableWRCVN specimens had differentproperties from standard AWS CVNspecimens. The pattern of resultsfrom WRCVN tests was shiftedapproximately 20 C to 40 C (35 F to70 F) higher then the standard AWSCVN specimens. The WRCVN specimens should reflect Fillet weldproperties more accurately becausethey are taken from the root of what isin essence a multiple-pass Fillet the pattern seen among the weather-ing specimens can be extrapolated toother consumables, then the standardtest overestimates weld (1998) demonstrates thatstandard AWS CVN test results arenot representative of productiongroove welds, and typically havehigher toughness values than produc-tion welds.

9 Fillet welds differ evenmore from the standard test weld , andso are even less likely to be ad-equately represented by the 1: Welding parametersConsumableDesignationFlux ElectrodeHeat Inputs(kJ/in)LowHighNeutral Flux960L-613/32" Flux780L-615/64" " Summary Report 1501-S 3 .The results of this researchindicate that the tensile strengthrequirements of the weld certificationtests are adequate to ensure thestrength of Fillet welds. Based uponthese results, the weld qualificationtests presently required in AWS not necessary to ensure thestrength of a Fillet weld . The T-bendspecimen did not provide a usefulmeasure of the strength or ductility ofa Fillet weld . The T-weldment doesprovide a simple means to evaluatethe influence of double-sided weldupon the geometry of the weld andmelt-through of the stem. TheWRCVN specimen provides aconvenient method of characterizingthe toughness of the Fillet weld rootmaterial. The WRCVN toughnessmay be comparable to or less than thetoughness measured in the standardall- weld metal tests.

10 The WRCVN test is recommended as a simplemeans to ensure that the Fillet weldtoughness is adequate. Toughnesscomparable to the base metal shouldbe sufficient for the root of the base metal is directly adjacent tothe weld metal at the root of the , a fracture will propa-gate in either the weld or base metal,whichever has the lowest is no benefit to having the weldmetal toughness significantly tougherthan the base metal. A weld roottoughness corresponding to the non-fracture-critical base metal require-ment for 4-inch plates in TemperatureZone III should be adequate for allbridges. For example, the requiredtoughness for Gr. 50 steel would be20 ft-lbs at 10o F, per ASTM A 709 Table upon the results of thisstudy, the following recommendedchanges to the specifications areproposed:1. The consumable supplier shallperform the following testsannually:a. Two weld certification tests,one at the highest and the otherat the lowest weld heat inputrecommended by the manufac-turer.