Plasma Transferred Arc Repair Welding of the Nickel-Base Superalloy IN-738LC

Transcription

1 JMEPEG (1997) 6: International Plasa Transferred Arc Repair Welding of the Nickel-Base Superalloy IN-738LC C.Y. Su, C.P. Chou, B.C. Wu, and W.C. Lih Plasa transferred arc welding (PTA) has been considered a proising process to restore worn areas of land-based gas turbine blades and vanes. The objective of this investigation was to study the effect of PTA welding on the repairing of IN-738LC superalloy coponents. Tensile tests were conducted on speciens welded with various cobinations of paraeters. Roo teperature, 760 ~ and 980 ~ were selected as tensile test teperatures. High-teperature phase transfored, during solidification, were identified by differential theral analysis (DTA). The weld-pool shapes and icrostructures of welded speciens prepared by various welding paraeters were evaluated by optical etallography (OM), a scanning electron icroscope (SEM) equipped with energy dispersive x-ray spectroeter (EDS), and icrohardness testing. Results of this study showed that PTA welded speciens exhibited 96% noinal tensile strength of IN- 738LC base aterials. Specien failure was observed predoinantly in the base aterials instead of in the heat-affected zone (HAZ) for gas tungsten are weld (GTAW) repair weldents. IN-738LC is considered susceptible to weld cracking during fusion welding; however, using a low-input heat repair welding process (PTA), cracking susceptibility could be iniized by the optiized welding paraeters. i Keywords 1. Introduction IN-738LC, Ni-base alloy, plasa, welding IN-738LC, which is a y precipitation-hardening nickelbase superalloy, has been extensively used as land-based gas turbine blade/vane aterials. In general, superalloys were considered difficult to weld, especially those exhibiting precipitation-hardening reactions involving titaniu and/or aluinu (Ref 1). Microfissures have been observed to occur in the welded etal and the heataffected zone (HAZ) during welding and during the subsequent postweld heat treatent (Ref 1-4). Heat-affected zone cracking was attributed to liquation cracking and a subsolidus ductility dip (Ref 5-7). In the first type, cracking is proposed to occur during the final stages of solidification of the thin liquid fils fored on HAZ grain boundaries. Two different echaniss have been proposed for the origin of the liquid fil on the grain boundaries: (1) grain-boundary elting caused by local copositional variations arising fro segregation and (2) liquid fil foration on grain boundaries caused by the constitutional liquation of priary carbides and carbonitrides, borides, sulfides, and so forth. It has been suggested that the second type of HAZ cracking occurs in the solid state exhibiting a drop in ductility below the solidus teperature. It is shown in Fig. 1 (Ref 8), indicating that the tendency of strain-age cracking depends on the total titaniu and aluinu content for several nickel-base superalloys. According to Prager's study (Ref 8), the weldability of the IN-738LC is poor and not suitable for fusion welding. Jahnke (Ref 9) eployed high preheat teperature (1120 ~ of electron bea (EB) C.Y. Su and C.P. Choo, Departent of Mechanical Engineering, National Chiao Tung University, Hsinchu, Taiwan, R.O.C.; and B.C. Wu and W.C. Lih, Industrial Technology Research Institute, Materials Research Laboratories, Hsinchu, Taiwan, R.O.C. welding to reduce the icrofissuring of postweld IN-738LC. Lowering the welding energy input ay decrease the tendency of cracking. The lower heat input reduces the volue of weld etal that undergoes shrinkage during cooling, therefore reduces the theral strain, and consequently decreases the cracking sensitivity. Electron bea welding exhibited very low heat inputs; however, this process ust proceed in high vacuu and is generally not econoical. The purpose of this study was to apply plasa transferred arc welding (PTA), which was considered an energy con- a 4 l! = 3,it Q7t3C Q Hat Id 200 AFI IOA Q r Udlet 700 Udtot $OOQ ~) 0 Astroloy /-IN73~LC -,,,,,, ~. ~ 5001 // / '. /OUn,tep 1753 [Readily WlldablwJ '~'~'~'@lllno' 41 O Waspalo)' Rene' 62/,.7,r I 750 TM'~. ~i~ll 252 ~ ~ (S)tnconel 716 ~'~inconel I ~'~ O IN 100 I 1 I I I I I I 1 I ~" ",J lltlntu, w/0 Fig. 1 Prager's plot of weldability versus aluinu and titaniu content, with several current cast nickel-base superalloys added, based on strain-age cracking susceptibility (Ref 8) Journal of Materials Engineering and Perforance Volue 6(5) October

2 stricted and concentrated process, on IN-738LC plates to result in lower heat inputs. The filler etal of PTA is in powder for and was added directly into PTA torch. As a result, lower dilution and high deposition rate were also expected as the advantages of PTA process for IN-738LC coponents repairs. 2. Materials and Experiental Procedures The base etal used in this study was IN-738LC nickel-base superalloy. The copositions of the alloy were deterined using inductively coupled plasa-atoic eission spectroscopy Table 1 Copositions of IN-738LC by ICP-AES detection Coposition, wt% Cr Co AI Ti AI + Ti Mo W Ta Nb C B Ni Base etal (a) (a) Bal PRAXAIR (a) (a) Bal NI (a) Did not detect Table 2 Welding paraeters Specien Current, Voltage, Travel speed, Powder feed rate, No.(a) A V /in g/in RWHT GEl_ yes GEI_ yes GEI_ yes GE2_ yes GE2_ yes GE2_ yes GE3_ no GE no GE no GE4_ , no GF_,4_ no GE4_ no DEI_ yes DEI_ yes DEI_ yes DE3_ no DE3.._ no DE3_ no Note: arc length: 10 ra, electrode diaeter (tungsten): in. (3 ) noinal, plasa gas fluid rate: 3 ft3/h (0.023 L/s), protective gas: argon, preweld heat treatent (RWHT): 1120 ~ h, AC, postweld heat treatent (PWHT): 1120 ~ h, AC and 845 ~ h, AC. (a) GE is the joint weld; DE is the deposition weld. L Welding t Direcd~ I (a) (n) Fig. 2 The torch echanis ofpta (Ref 10) Fig. 3 Proposed (Ref 12) surface and subsurface fluid flow in the weld pool. A, surface tension teperature coefficient negative; B, positive surface tension teperature coefficient Volue 6(5) October 1997 Journal of Materials Engineering and Perforance

3 (ICP-AES) and is shown in Table 1. The IN-738LC plates were sliced fro daaged Asea-Brown-Boveri (ABB) land-based gas turbine vanes root with 3 thickness. PRAXAIR NI (Praxair Surface Technologies, Inc., Indianapolis, IN) was selected as feed powder. The PTA syste used was Starweld 3 by Stellite Copany with Model 600 Torch. In order to investigate the potential of repairing blades using a PTA syste (Fig. 2), two groups of speciens were welded as follows: (1) direct deposited weld designed to siulate hot corrosion daage of blade surface, and hence restore the original blades and (2) null gap joint weld designed to siulate the penetrated cracking of echanical stress. The paraeters of this experient are given in Table 2. The deposited rate of filler etal was fixed. Prior to welding, the saples were either in the as-cut condition for the root of the daaged blades and in the solutiontreated condition as shown in Table 2. Postweld heat treatent was perfored on all welded saples (Ref 11) o~ v E o --I 60 c t- ~ a 50 HT 0.20 RWHT 9 RWHT , I, I, I i I, Input Energy Density (J/) , I, I, I, I, Input Energy Density (J/ra) (a) (b) o~ 40 t- O ",, o :~ 0.25 n" a 30 t RWHT WriT D 0.15 RWHT WHT (c) I = I = I = Input Energy Density (J/) (d) i I i I i I i Input Energy Density (J/ra) Fig. 4 (a) Weld dilution versus weld input energy density for the non-rwht and RWHT speciens in the joint type. (b) Weld depth/width ratio versus weld input energy density for the non-rwht and RWHT speciens in the joint type. (c) Weld dilution versus weld input energy density for the non-rwht and RWHT speciens in the deposition type. (d) Weld depth/width ratio versus weld input energy density for the non-rwht and RWHT speciens in the deposition type Journal of Materials Engineering and Perforance Volue 6(5) October

4 Fig. 5 The 7' distribution in base aterials with different heat treatent (SEM). a, coarse 7'; b, cooling 7'. (a) non-rwht and non-pwht. (b) non-rwht and PWHT. (c) RWHT and non-pwht. (d) RWHT and PWHT Differential theral analysis (DTA) test was perfored using a DuPont Theral Analyzer 2000 (TA Instruent, Inc., Castle, DE) with epty aluina cup as a reference. Saples of as-received aterial, approxiately 40 g, were heated to 1500 ~ at a rate of 0.33 ~ during heating and cooling under flowing high-purity argon while solidification data were recorded. Tensile test speciens were achined fro welded saple according to ASTM E 8M. Tensile tests were perfored at roo teperature, 760 ~ and 980 ~ with a strain rate of 0.04 /s. The electrolytic etchant used for IN-738LC was 10 L H3PO 4, 45 L H2SO 4, and 45 L HNO 3 at 6 Vdc for revealing the icrostructure. Metallographic observation was perfored in the fusion zone, the HAZ, and the base aterials. An iage analyzer was used to calculate the dilution rate and depth/width ratio of the fusion zone. Scanning electron icroscopy was used to easure 7' size and distribution. A JEOL 840 scanning electron icroscope (JEOL Ltd., Tokyo) equipped with an OXFORD energy-dispersive spectroscopy (EDS) syste (Oxford Instruents Microanalysis Group, High Wycobe Bucks, England) with a LINK detector window (Oxford Instruents Microanalysis Group, High Wycobe Bucks, England) enabled a qdalitative analy sis of light eleents such as carbon and nitrogen. Volue fraction of icrostructural constituents were easured using an OPTIMUM Iage Analyzer(BioScan,Incorporated,Edonds,WA). 3. Results and Discussion 3.1 Effect of Welding Paraeters Shallow welds are observed for any alloys because in these aterials the surface tension has a negative teperature coefficient. The surface tension gradient drives fluid flow fro the center of the pool outward, as illustrated in Fig. 3(a). The outward fluid flow provides substantially greater heat transport to the outside of the pool. Figure 3(b) shows the surface tension teperature coefficient is positive. Weld cross-sectional area, width, and depth of weld bead were easured by iage analyzer. The effect of input energy density on the dilution and depth/width ratio for both the joint type and the deposition type are shown in Fig. 4. It is shown that both dilution and depth/width ratio increased as input energy density increased for both types. That is, weld depth increased 622--Volue 6(5) October 1997 Journal of Materials Engineering and Perforance

5 Fig. 6 The T-T' eutectic phase of base aterial precipitated in grain boundary after high-teperature exposure for a long period (SEM) (b) Fig. 7 Point carbide existed at dendritic intergranular region of as-welded etal. Welding current: 90 A, travel speed: 240 /in. (a) Light icrograph. 800 x. (b) SEM-EDS analyses of the carbide in the dendrite intergranular pairability. On the other hand, for non-rwht IN-738LC coponents, weldability ay be susceptible to the degraded icrostructure caused by long-ter high-teperature operation, and this is discussed in section 3.2. Fig, 8 DTA therogra for filler etal IN-738LC faster than weld width. It was proposed that the positive surface tension teperature coefficient doinates the flow odel of filler aterials in the weld (Ref 12). It was also proved that porosity ight be iniized when this flow odel takes place in the weld (Ref 13). This echanis was also proved by icrostructural investigation of PTA weld in this study, which showed that porosity was inial. For the joint specien that was not heat treated before welding, the depth/width ratio decreased as the input energy density exceeded J/. This is indicated in Fig. 4(b), and it can be proved that the surface tension teperature coefficient turns to negative. The results of Fig. 4(b) indicate that preweld heat treatent (RWHT) is iportant to keep the depth/width ratio within a controlled range that ay benefit the weldability/re- 3.2 MetaUographic Investigation The size and distribution of y' phase in RWHT IN-738LC speciens are shown in Fig. 5. Coarse y' precipitate occurred after long-ter high-teperature operation, It was observed in Fig. 5(c) that fine-cooled 1/precipitated at the atrix of coarse 7' for RWHT specien. For postweld heat treatent (PWHT) speciens, coarse t", precipitated y', and cooled y' were all observed as presented in Fig. 5(b) and (d). The ajor purpose of RWHT is to release the accuulated stress and redissolve the segregation in IN-738LC cast after long-ter operation and hence hoogenize the atrix of IN-738LC to iniize the possibility of welding hot cracking. As shown in Fig. 6,1(-1" eutectic phase generally segregates at grain boundaries. Hot cracking and strain-aging cracking during repair welding are easily induced in this region due to the high hardness character of y-y' eutectics. Figure 7(a) shows the icrostructure of RWHT welded etal for which welding current was 90 A, travel speed of 240 Journal of Materials Engineering and Perforance Volue 6(5) October

6 rrdin, and powder feed rate of 8 g/in. The black spotty precipitates around the dendrites observed in Fig. 7(a) were identified as carbides with tantalu, olybdenu, and titaniu as ain constituents. Figure 8 shows the DTA therogras of IN-738LC filler powder. The solidus teperature 1456 ~ was identified by the initial deviation fro the local baseline, and the liquidus teperature was deterined as 1481 ~ by the peaks on therogra. There is at least one inor constituent solidification also identified at 1413 ~ in DTA therogra. Figure 9 shows the icrostructures of the PTA IN-738LC welds. For non-rwht and non-pwht welds, y' precipitates were not observed; however, for both non-rwht/pwht and RWHT/PWHT welds y' precipitates were observed in a hoogeneous distribution. The effects of RWHT and PWHT treatents on the hardness of both joint and deposition welds are shown in Fig. 10(a) and (b), respectively. The weld hardness is coparatively higher for joint RWHT speciens than the hardness for joint non-rwht speciens as indicated in Fig. 10(a). This is tentatively to be inferred as the result of discontinuous carbides precipitated at finer dendrites of RWHT speciens, whereas the discontinuous carbides for non-rwht speciens precipitated with coarser dendrites. Siilar results of weld hardness were observed for deposition speciens, as shown in Fig. 10(b). Liquation cracking is believed (Ref 13) to be associated with local or partial elting in the grain boundary adjacent to the fusion line in the HAZ. As shown in Fig. 11, partial elting was observed in the HAZ and believed to induce the icrofissure fro HAZ to the dendrite boundaries of fusion zone. Solute segregation at grain boundaries plays an iportant role in the type of localized elting as shown in Fig. 11. The distribution off precipitates in the HAZ is shown in Fig. 12. For non-rwht/non- PWHT specien, "/in HAZ were coarsened and hence coalesced. However, for RWHT/PWHT specien, coarsened y', cooled y', and age precipitated y' are observed incidentally as shown in Fig. 12(b). This indicates that part of the HAZ was reelted during welding and new y' phase precipitated during cooling. Weld cracking was not observed in this study due to the choice of low current/tfigh travel speed of PTA torch paraeters. 3.3 Mechanical Property The results of tensile tests are shown in Fig. 13. At 760 and 980 ~ joint PTA speciens exhibited siilar ultiate tensile strengths as base etal. However, for tensile tests at roo teperature, the joint PTA specien exhibited 80% of the ultiate tensile strength of the base etal. Fracture failure was observed predoinantly in the base aterials and occasionally occur in the HAZ. Conclusions Fig. 9 The y' distribution in the fusion zone with different heat treatent (SEM). Welding current: 90 A, travel speed: 240 /in. (a) non-rwht and non-pwht. (b) non-rwht and PWHT. (c) RWHT and PWHT Using plasa transferred arc welding, cast nickel-base superalloys, especially IN-738LC, can be welded without deleterious HAZ or weld etal cracking. Weld cross-sectional analysis showed that preweld heat treatent could iprove the weldability of joint IN-738LC coponents by long-ter high-teperature operation. Metallographic inspection showed a very fine dendritic solidification structure. Using proper postweld heat treatent, the distribution of y' precipitates can becoe hoogeneous Volue 6(5) October 1997 Journal of Materials Engineering and Perforance

7 550 [., ~,----/k--t3e1 non-rwht non-pwht ;> r~ base etal... "77 ii'7;...,ik (~1 non-rwht PWHT (~33 RWHT non- PWHT ----II--- (~3 RWHT PWHT I --- f u s i o n zone 350 distance fro centerline () (a) I ----/k-- DE1 non-rwht non-pwht J- DE3 RWHT non- PWHT DE1 non-rwht PWHT --l--- DE3 RWHT PWHT 350 I I I I I I f I I I I I ~5 ~ eq Distance () (b) Fig. I0 (a) Microhardness distribution curve across PTA fusion zone in the joint type. Welding paraeters: current, 115 A; voltage, 25.3 V; speed, 240 /in; powder rate, 8 gw/in. Hardness test load is 200 g. (b) Microhardness distribution curve across PTA fusion zone in the deposition type. Welding paraeters: current, 50 A; voltage, 21.3 V; speed, 240 /in; powder rate, 8 gw/in. Hardness test load is 200g. Journal of Materials Engineering and Perforance Volue 6(5) October

8 Fig. 11 Liquation cracking of fusion zone induced fro the partial elting zone, welding current: 90 A, welding speed: 240 /in, non- RWHT (OM) Fig. 12 The T' distribution of HAZ was differently heat treated (SEM). Welding current: 90 A, travel speed: 240 /in (Ref 11). (a) non-rwht and non-pwht, (b) non-rwht and PWHT. (c) RWHT and PWHT Fig, 13 Ultiate tensile strength (UTS) of PTA-welded IN-738LC at 25,760 and 980 ~ (UTS of base aterial was referred fro Ref 11) Volue 6(5) October 1997 Journal of Materials Engineering and Perforance

9 The results of tensile tests showed that PTA welded speciens exhibited 96% noinal ultiate tensile strength of IN-738LC base aterials. Specien failure was observed predoinantly in the base aterials. References 1. C.T. Sis, N.S. Stoloff, and W.C. Hagel, Superalloys ll, John Wiley & Sons, D. McKeown, Weld. J., Vol 3, 1971, p 201s 3. K.C. Wu and R.E. Herfert, Weld. J., Vol 1, 1967, p 32s 4. W.P. Hughes and T.F. Berry, WeldJ., Vol 8, 1967, p 361s 5. E.G. Thopson, Weld.J., Vo148 (No. 2), 1969, p 705s 6. M.J. Flether, Weld. Met. Fabr., Vol 38 (No. 2), 1970, p W.A. Owczarski, D.S. Duvall, and C.P. Sullivan, Weld. J., No. 4, 1966, p 145s 8. M. Prager and C.S. Shira, Weld. Res. Counc. Bull., No. 128, B. Jahnke, Weld. J., No. 11, 1982, p 343s 10. PTA Training Manual, Deloro Stellite, Aerospace Structure Metals Handbook (V), Code 4217, Metals and Ceraics Inforation Center Battelle Colubus Laboratories, P. Burgardt and C.R. Heiple, Weld. J., No. 6, 1986, p 150s 13. S. Kuo, Welding Metallurgy, John Wiley & Sons, 1987, p 102 Journal of Materials Engineering and Perforance Volue 6(5) October