V. Verderaime* and R. Vaughan*NASA Marshall Space Flight Center, Huntsville, Alabama 35812Though butt-welds are among the most preferred joining methods in aerostructur_ their strength dependenceon inelastic mechanics is generally the least understood. This study investigated experimental strain distributionsacross a thick aluminum U-grooved weld and identified two weld process considerations for improving the multipnssweld strength. One is the source ofpeaking inwhich the extreme thermal expansion and contraction gradient ofthefusion heat input across the groove tab thickness produces severe angular distortion that induces bending underuniaxial loading. The other is the filler strain hardening decreasing with increasing filler pass sequences, producingthe weakest welds on the last weld pass side. Both phenomena are governed by weld pass sequences. Many industrialwelding schedules unknowingly compound these effects, which reduce the weld strength. A depeaking index modelwas developed to select filler pass thickness, pass numbers, and sequences to improve depeaking in the weldingprocess. The result was to select the number and sequence of weld passes to reverse the peaking angle such as tocombine the strongest weld pass side with the peaking induced bending tension component side to provide a moreuniform stress and stronger weld under axial tensile loading.NomenclatureE = elastic modulus, ksiF = material strength, ksiH = specimen thickness, in.h = weld pass thickness, in.K = inelastic strength coefficient, ksiM = induced moment, in.-kipsm = weld sequence numberN = applied axial load, kipsn = strain-hardening exponent, total number of weld passesT = temperature, °Ft = U-groove tabs thickness, in.w = specimen width, in.a = coefficient of thermal expansion, in./in./°F_b = peaking angle, radSubscriptse = elastic variablei = strain gauge number, weld pass seriesj = weld pass numberk = designated temperM = moment variableN = axial load variablep = inelastic variabletu = tensile ultimatety = tensile yieldct = thermal variableIntroductionSweld structuralthicknessesenvironmentsincrease, andweldcomponentdevelopmentsizes increase,andprocessesbutt-become more complex, and joint strengths axe less predictable. Oneearly study modeled a uniaxial butt-weld specimen having differentinelastic lateral contraction rates between preweld material and weldfiller and discovered a metallurgical discontinuity at the interfaces.lDiscontinuity stresses, especially transverse shear, were later ex-perimentally verified on a thick weld cross section in uniaxial test. 2The primary objective of the reported study was to further exploreReceived Feb. 15, 1996; revision received Sept. 25, 1996; accepted forpublication Sept. 30, 1996. Copyright _ 1996 by the American Institute ofAeronautics and Astronautics, Inc. No copyright is asserted in the UnitedStates under Title 17, U.S. Code. The U.S. Government has a royalty-freelicense to exercise all rights under the copyright claimed herein for Govern-mental purposes. All other rights are reserved by the copyright owner.*Aerospace Engineer, Structures and Dynamics Laboratory.133the multipass welding process and resulting structural properties ofweld filler passes from experimental test data 3 and to identify weldprocess variables that should improve strength performance.Weld Peaking SpecimenThe aluminum test specimen shown in Fig. 1 was a 0.71 in. wideslice from a double U-grooved butt weldment of two very large ma-chined 2219 aluminum panels. The weld filler was 2319 aluminumwith the beads ground off. The panels were 1.4 in. thick, and thebutted tab thickness between the double U grooves was 0.375 in. Itwas tungsten inert gas welded using the symmetrical welding sched-ule noted in Fig. 1, referred to as a normal welding schedule. Thebutted tabs were tacked and then continuous fusion welded from thesame side, incurring a net initial peaking angle _. Weld peakingis an unintentional angular panel displacement resulting from weldthermal gradient strain. Subsequent welds were fleer passes seriallyapplied, first in the groove opposite passes 1 and 2 and then on thereverse side groove, for a total of eight passes.Weld pass 1 (Fig. 1) was crucial to the butted edge mismatch. Inthis weld pass, the double U-groove tabs at the midplane were butted,the panel surface planes were aligned, the assembly was constrained,and the butted tabs were fusion tack welded (without filler material)on one side. The tack weld pass produced local thermal expansion onthe butted tabs and was followed by cooling contraction. The coolinginduced a tensile strain on the tack weld side and compression onthe unfused side of the tab, which mildly peaked the panels with theobtuse angle on the tack welded (pass 1) side.The intense weld heat input from pass 2 severely increased theadverse peaking angle. It was another fusion weld pass applied con-tinuously on the same side of the tack weld, and it had the highestheat input rate to fuse the total tab thickness. The associated extremethermal expansion and contraction gradient across the tab thicknessproduced the maximum peaking angle in the process with the obtuseangle again on the heat source (pass 1) side.The next three passes were weld filler passes (thinner than thetabs) requiring less heat and were applied in the groove opposite the:..._ w _ _ i fidplane _loa dFi_ 1 Test specimen configuration.
Read full abstract
Oct 8, 2024
Kelvin F Long 
Open Access
