Preliminary Investigation On Real Time Induction Heating Assisted .

Zhang Supplement January 2015 Layout 1 12/11/14 2:15 PM Page 10WELDING RESEARCHABCDEFig. 2 — Captured images of arc shape: A — Without induction heating; B — L 5 mm,250 V; C — L 5 mm, 450 V; D — L 20 mm, 250 V; E — L 20 mm, 450 V.Table 2 — Optimized Welding ParametersWelding Voltage (V)26Welding Current (A)Welding Speed (mm/min) Water Depth (mm)160shape with parameter L and induction heating voltages. Due to optimum parameters and flux-cored wire,the welding arc was steady during theunderwater welding process withoutinduction heating (Fig. 2A). While theinduction coil was installed, the arcstability was reduced. It was observedthat the parameter L played a majorrole in arc stability. When the parameter L was 5 mm, the welding arc wasextremely unstable and even arc interruption appeared in Fig. 3A. At thesame time, when parameter L was increased to 20 mm, the welding arcshape was stable. Therefore, a continuous and uniform weld could be observed in Fig. 3B. Welding discontinuities, such as incomplete fusion andundercut, were not found. In addition, the induction heating voltage affected arc stability and the arc stability decreased with increased voltage.To investigate the influence of voltage on the joint, the parameter L wasfixed as 20 mm in the subsequentexperiments.Cross Section MacrographsQ460 sheets were underwater weld-145300ed at a fixed welding parameter (Table2) and at various induction heatingvoltages ranging from 250 to 450 V inorder to clarify the effect of inductionvoltage on weld penetration. Crosssectional macrographs of the jointswith different voltages are shown inFig. 4. According to the results, weldpenetration and HAZ increased withthe increasing voltage. As was known,the width of the HAZ depended primarily on heat input. The heat inputwas the sum of welding heat input andinduction heating. Therefore, the effect of induction heating was equal toincreasing the welding heat input. Inaddition, the induction heating madethe temperature field of the weld zonerelatively more uniform.Microstructure Characteristicsof the JointsThe HAZ for Q460 delivered asrolled sections mainly consisted of twodistinct zones: coarse-grained HAZ(CGHAZ) and fine-grained HAZ (FGHAZ). Typical HAZ temperature vs.time profiles during the underwaterwet welding are shown in Fig. 5. Ac-10-s WELDING JOURNAL / JANUARY 2015, VOL. 94cording to the results, the cooling ratein wet welding was extremely higherthan in air welding. For instance, thecooling rate of the temperature rangefrom 800 to 500 C could rise sharplyto 100 C/s, which was more than thecritical cooling speed of martensiteformation.Figure 6 showed the optical microstructure of the weld zone withvarious induction heating voltages.Based on the theory of welding metallurgy, as the austenite phase wascooled down from high temperature,ferrite nucleated at the grain boundary at 770 –680 and then grew inward. This ferrite was proeutectoidferrite (PF), which is also called grainboundary ferrite (GBF). When thetemperature dropped to 500 , thetransformation of acicular ferrite(AF) occurred. The acicular ferritephase was a desirable phase becauseof the excellent plasticity and toughness characteristics (Ref. 25). As thecooling rate increased, the transformation product changed to bainiteand martensite phase and reducedthe mechanical properties. The microstructure of the weld zone in airwelding was composed of proeutectoid ferrite and acicular ferrite phase.As was mentioned previously, the acicular ferrite phase had excellent plasticity and toughness, due to the interlocking nature of the acicular ferriteand the fine granular size. Therefore,the mechanical properties were satisfactory. Compared to air welding, themicrostructure of underwater welding was a mixture of lath martensite,upper bainite, and proeutectoid ferrite — Fig. 6A. The bainite sheaf andmartensite lath nucleated and grewfrom prior austenite granular boundaries. The formation of lath martensite and upper bainite were detrimental to the weld properties, owing tothe easy crack propagation paths. Asthe induction heating voltages increased (Fig. 6 B–D), the volume fraction of lath martensite and upper bainite decreased with ferrite phases increasing. Moreover, the lath martensite and upper bainite phase disappeared as the voltage reached 350 V.The transformation product (Fig. 6C)could change from upper bainite andlath martensite to acicular ferrite andproeutectoid ferrite. Therefore, themicrostructure of the weld metal was

Zhang Supplement January 2015 Layout 1 12/11/14 2:15 PM Page 11WELDING RESEARCHAABCDBFig. 3 — Weld appearances: A — L 5 mm, 250 V; B — L 20 mm, 250 V.Fig. 4 — Cross sectional macrograph of the joints: A — 0 V; B — 250 V; C— 350 V; D — 450 V.ABDFig. 5 — Measured temperature vs. time curves of under water wet welding without induction heating.similar to air weld, the only differencewas the morphology of the proeutectoid ferrite. Increased acicular ferritecontent in the microstructure improved cracking resistance, while upper bainite and lath martensite deteriorated the mechanical properties ofthe joint. The dimension of theproeutectoid ferrite was increasedwith the increase in the voltage. Asthe voltage was 450 V (Fig. 6D), themorphology of the proeutectoid ferrite was coarsening and a ferrite sideplate (FSP) was found. Because of thelimited output voltage of the induction heating system, the inductionheating was higher than 450 V, andthe microstructural evolution andmechanism are to be investigated inthe future.To understand the mechanism ofweld microstructural evolution, temperature vs. time profiles of differentinduction heating voltages without being subjected to welding are shown inFig. 7. The parameters of t8/5, for a giv-CEFig. 6 — The optical microstructure of the weld zone with various inductionheating voltages: A — 0 V; B — 250 V; C — 350 V; D — 450 V; E — in air.en hardenability steel, determined thehardenability of the transformationproducts, which should be taken intoconsideration to investigate the effecton susceptibility to hydrogen-inducedcracking. The data of temperature vs.time curves are shown in Table 3. Asthe induction heating voltage was 250V, the Tmax reached 412 C. Therefore,the microstructure of Fig. 6B was similar to that shown in Fig. 6A becausethe t8/5 determined the transformationproducts. As the induction heatingvoltage was 350 V, the Tmax was increased to 609 C and the t8/5 was prolonged at the same time. Therefore,the transformation products changedfrom upper bainite and lath martensite to acicular ferrite and proeutectoid ferrite due to the fact t8/5 was prolonged. A comparison of temperaturecurves of 0 and 450 V is shown in Fig.8. It could be seen that the prolongation of t8/5 was extremely obvious.That’s the reason for the evolution ofthe microstructure of the weld metal.Figure 9 showed the optical microstructure of the partially meltedzone with and without inductionheating. The red line was the weld interface of the joint. It could be seenthat lath martensite (M) and coarsening Widmanstätten (W) structure waspredominant in the coarse-grainedHAZ in Fig. 9A. The ferrite phase precipitated first in the coarse-grainedaustenite grain boundary, and thengrew into the austenite in the form ofreticular structure (also called Wstructure), resulting in splitting thematrix structure, even generating thecrack. And the lath martensite composed of vast coarse lath was beneficial for crack initiation and propagation. Therefore, the mechanical properties of the joint were decreased.However, as the voltage was 350 V,granular bainite was predominantand grain coarsening was relieved.Thus, the tendency to crack was decreased, and the mechanical properties of the joint were increased.JANUARY 2015 / WELDING JOURNAL 11-s

Zhang Supplement January 2015 Layout 1 12/11/14 2:15 PM Page 12WELDING RESEARCHFig. 8 — Compared temperature vs. time curves during un derwater wet welding with various induction heating volt ages.Fig. 7 — Temperature vs. time profiles of different induc tion heating voltages without being subjected to welding.BAFig. 9 — The optical microstructure of the weld interface: A — 0 V; B — 350 V.Table 3 — The Data of Temperature vs. Time CurvesInduction Heating Voltages (V)250350450Mechanical PropertiesTensile Testing and FractureMorphologyFive prepared tensile specimensfrom each joint were performed usinga fully computerized tensile testingmachine with a loading rate of 1mm/min at room temperature to evaluate the influence of various inductionheating voltages on the mechanicalproperties of the joint. The geometryof the tensile specimens and tensilestrength vs. voltage curves are shownin Fig. 10. The thickness of the specimen was 4 mm. The tensile strengthTmax( C)t8/5 (s)41260971202435of specimens without induction heating was 444 MPa, about 82.2% of thebase metal (540 MPa). The tensileproperty increased gradually with increasing voltages. As the voltage was450 V, the tensile strength reached532 MPa, about 98.5%, and all thejoints fractured roughly in the HAZ.Joint efficiency increased from 82.2 to98.5%.In order to observe the fracturemechanism, SEM was carried out to analyze the fracture morphology. Figure11 shows the typical fracture surface ofspecimens with different voltages. Itcan be seen that a quasi-cleavage fracture mode was dominant in Fig. 11A,due to plenty of cleavage plane appear-12-s WELDING JOURNAL / JANUARY 2015, VOL. 94Fig. 10 — Tensile strength vs. inductionheating voltage curves.ance. The size of the cleavage plane wasrelated to the crack path. The largecleavage planes demonstrated very lowcrack propagation energy, while thesmall cleavage plane exhibited highercrack propagation energy (Ref. 26). Theformation of lath martensite and Wstructure in the HAZ was detrimental totensile property due to the easy crackpropagation paths. Once the crack occurred during the tensile test, it couldpropagate along the paths of lathmartensite and W structure rapidly.Therefore, tensile strength without induction heating was the lowest. Thedimple characteristics became predominant as the voltage was increased in Fig.11B–D. Cleavage planes were in smallproportion while dimples were in largeproportion, as shown in Fig. 11B.Therefore, the fracture morphology hadthe characteristic of ductile fracture.While the voltage was above 350 V, thecleavage planes disappeared, instead ofuniform dimples, which was the typicalfeature of ductility. Nonmetallic inclu-

Zhang Supplement January 2015 Layout 1 12/11/14 2:15 PM Page 13WELDING RESEARCHABCDFig. 12 — Impact energy vs. induction heating voltagecurves.ABFig. 11 — SEM images of the fractured surface of specimens withvarious voltages: A — 0 V; B — 250 V; C — 350 V; D — 450 V.sions phase were disorderly distributedin the inter-tear edges. Hence, tensileproperty of the joints improved toabout 98.9% of the base metal.Impact Testing and FractureMorphologyCharpy V-notch impact tests wereconducted at 20 C on an instrumenteddrop weight impact tester. The specimens were extracted in the weld’s perpendicular direction from the middlethickness of the as-welded specimenwith notches positioned at the centerof the weld metal. The impact energyof the joints was the average of fivespecimens. The geometry of theCharpy impact V-notch specimens andimpact energy vs. induction heatingvoltage curves are shown in Fig. 12.The thickness of the specimen was 10mm. The impact energy value of specimens without induction heating was36 J. As the voltage increased, the impact energy increased consistently. Finally, as the voltage was 350 V, the impact energy could reach 68 J. The results suggested that induction heatingcould increase the impact propertiesand the toughness of the joint.For the purpose of observing thefracture mechanism, SEM was used toanalyze the fractureCmorphology. Figure 13shows the SEM micrographs of impact fracture surface morphologies for different voltages. It can be seen thata cleavage fracture modeis dominant in Fig. 13A,due to a network ofFig. 13 — SEM images of the fractured surface of specimenscleavage steps known aswith various voltages: A — 0 V; B — 250 V; C — 350 V.a river pattern. Cleavagewas a low-energy fracpresent in large proportion. The apexture that propagated along wellof the fan points back to the fracturedefined low-index crystallographicorigin. While the voltage was 350 V,planes known as cleavage planes. Thethe cleavage planes disappeared, inbranches of the river pattern joinedstead of uniform dimples, which wasin the direction of crack propagation.the characteristic of ductile fracture.Meanwhile, the formation of lathSome nonmetallic inclusions phasemartensite and upper bainite was adshowed in a disorderly distributionverse to the toughness of the jointand were surrounded by the interdue to the easy crack propagationtear edges. The reason for these repaths. Once the crack occurred, it rapsults was that the acicular ferrite actidly propagated in a straight lineed as the crack arrester and increasedalong the lath martensite and upperthe crack propagation energy. Therebainite paths. Therefore, the impactfore, the impact property of the jointsenergy value without induction heatwas greatly improved from 36 to 68 J.ing was the lowest. However, the dimples started to appear and were presBend Testingent in small proportion when thevoltage was at 250 V in Fig. 13B.Longitudinal three-point bend testsFeather markings, which are a fanwere conducted to measure the bendingshaped array of very fine cleavageductility at room temperature. The ansteps on a large cleavage facet, areJANUARY 2015 / WELDING JOURNAL 13-s

Zhang Supplement January 2015 Layout 1 12/11/14 2:15 PM Page 14WELDING RESEARCHFig. 15 — Schematic of Y slit self restrained cracking test (mm).Fig. 14 — Angle of bending vs. induction heating voltagecurves.gle of bending for the joints was the average of three specimens. The geometryof the specimens and angle of bendingvs. induction heating voltage curves areshown in Fig. 14. The thickness of thespecimen was 5 mm. According to theresults, the angle of bending values ofspecimens without induction heatingwas 21 deg, which indicated ductilitywas very low. At the same time, the angle of bending values increased rapidlywith increasing voltages. Finally, as thevoltage was 450 V, the angle of bendingvalues reached 88 deg. The results suggested joint ductility had beenincreased.Y Slit Restraint TestingBecause of the high quenching ratecaused by the water environment andbecause large quantities of hydrogen arepresent, hydrogen cracking is one of themost severe problems in the underwater welding of steel (Ref. 27). The cracking tests were carried out using a Y-slitrestraint test so that the intensity of therestraint could be related to actual fabricating conditions. The geometry ofspecimens subjected to Y-slit restrainttesting and the cracking ratio used as ameasure of the cracking susceptibilityare shown in Fig. 15. This is a ratio ofthe height from the root to the tip ofthe crack vs. the height from the root tothe surface of the weld metal.The cracking ratio vs. induction heating voltage curves are shown in Fig. 16.The carbon equivalent value of Q460steel was 0.6, which indicated the steelwas particularly sensitive to cracking,especially in underwater welding.Therefore, the cracking ratio of specimens without induction heating was82%. However, as the voltage reached150 and 250 V, the cracking ratio decreased rapidly to 45% and 22%, respectively. Finally, when the voltage wasabove 350 V, the cracking ratio reachedabout 10%. Typical weld cross sectionswith various induction heating voltagesare shown in Fig. 17. According to theresults, induction heating could reducethe cooling rate; therefore, the crackingsusceptibility was decreased.Microhardness ProfileVickers microhardness measurementacross the fusion zone was carried outwith a load of 100 g and load time of 10s. Results of hardness measurementsare shown in Fig. 18. The microhardness distribution indicated the microstructural characteristics of the joint.Increased hardness values of the weldmetal confirmed these microstructuralchanges. The location of the HAZ wasdetermined by metallographic observation and the hardness of the HAZ washigher than that of the weld metal. TheHAZ and weld metal hardness decreased with increased induction heating voltage. The maximum hardness ofHAZ without induction heating was 425HV, which was much harder than thatwith 250 and 450 V. The hardness values of the weld zone with 250 and 450V were relatively uniform because thelath martensite and upper bainite cotent decreased while the acicular ferriteand proeutectoid ferrite increased. Theresults indicated induction heating hada significant effect on the maximumhardness. The microhardness profile14-s WELDING JOURNAL / JANUARY 2015, VOL. 94across the weld indicated the microstructural characteristics of the joint.Induction heating made the joint microhardness relatively more uniform.Conclusion1) A novel real-time induction heating-assisted underwater wet weldingprocess was employed. The addition ofinduction heating could reduce the cooling rate of the joint in water environment to improve the microstructuraland mechanical properties of the joint.2) Arc stability was reduced with theaddition of induction heating. The parameter L played a major role in arc stability. As the parameter L increased to20 mm, the welding arc shape was stable. A continuous and uniform weldjoint could be observed.3) The content of martensite (M)and upper bainite (BU) phases decreased while the proeutectoid ferrite(PF) and acicular ferrite (AF) phases increased as the induction heating voltageincreased. Mechanical properties, suchas tensile, impact, and bending properties, increased as the induction heatingvoltages increased.4) Cracking was examined via a Y-slitrestraint test. The addition of inductionheating could decrease the cracking ratio from 82 to 10%. Therefore, induction heating could make cracking susceptibility decrease.References1. Brown, R. T., and Masubuchi, K. 1975.Fundamental research on underwater welding. Welding Journal 54(6): 178-s to 188-s.2. Rowe, M., and Liu, S. 2001. Recentdevelopments in underwater wet welding.Science and Technology of Welding and Join-

Zhang Supplement January 2015 Layout 1 12/11/14 2:15 PM Page 15WELDING RESEARCHABCDFig. 16 — Cracking ratio vs. induction heating voltagecurves.Fig. 17 — Typical underwater weld cross sections with various inductionheating voltages: A — 0 V; B — 150 V; C — 250 V; D — 350 V.cored arc welding. Journal ofMaterials Processing Technology213: 1370–1377.10. Maksimov, Y. S. 2010.Underwater arc welding ofhigher strength low-alloysteels. Welding International 24:449–454.11. Skorupa, A., and Bal, M.Fig. 18 — Hardness distribution along the joint.1996. The effect of aqueous environments on the quality of underwater‐welded joints. Weldinging 6: 387–396.International 10: 95–98.3. Łabanowski, J. 2011. Development12. Pope, A. M., Medeiros, R. C., andof underwater welding techniques. WeldingLiu, S. 1995. Solidification of underwaterInternational 25: 933–937.wet welds. Proceedings of the International4. Łabanowski, J., and Fydrych, D. RogalConference on Offshore Mechanics and Arcticski, G. 2008. Underwater welding — A reEngineering 3: 517–521.view. Advances in Materials Science 8: 11–22.13. Grubbs C. E. 1993. Underwater wet5. Ibarra, S. J., Reynolds, T. J., andwelding (a state of the art report). ProceedingGabriel, G. 1996. Amoco Trinidad selectsof the International Conference on Offshorewet welding repair option. Proceedings ofMechanics and Arctic Engineering 3: 111–118.the International Conference on Offshore Me14. Grubbs, C. E., Reynolds, T. J. 1998.chanics and Arctic Engineering 3: 109–112.State-of-the-art underwater wet welding.6. Yara, H., Makishi, Y., Kikuta, Y., andWorld Oil 219: 79s.Matsuda, F. 1987. Mechanical and metal15. Grubbs, C.E., and Reynolds, T. J.lurgical properties of an experimental cov1998. Underwater welding: Seeking highered electrode for wet underwater welding.quality at greater depths. Welding JournalWelding International 1: 835–839.77(9): 35–39.7. Liu, S., Olson, D. L., and Ibarra, S. J.16. Tsai, C. L., and Masubuchi, K. 1979.1995. Designing shielded metal a

A novel real time induction heating assisted underwater wet welding process was investigated. The addition of induction heating could reduce the cooling rate of the joint in underwater wet welding. The macro and microstructures, mechanical prop erties such as tensile, impact, and bending properties, and Y slit restraint testing were studied.