ROLE OF PHASE TRANSFORMATIONS IN RESIDUAL STRESS DEVELOPMENT IN MULTIPASS FERRITIC STEEL WELDS AND GLEEBLE TEST BARS

Transcription

1 ROLE OF PHASE TRANSFORMATIONS IN RESIDUAL STRESS DEVELOPMENT IN MULTIPASS FERRITIC STEEL WELDS AND GLEEBLE TEST BARS Steve Spooner, Stan A David, and C R Hubbard Oak Ridge National L a b o r a t o ~ Oak Ridge, TN 3783 Abstract Neutron strain scanning has proven very effective in nondestructive mapping of distribution of residual stresses in weldments Strain scanning of Gleeble test bars of 2 1/ Cr1 Mo steel has been carried out in conjunction with strain scanning investigations of a multipass weld in in plate of same alloy The residual stresses in Gleeble bars depend on time spent at maximum temperature and rate of cooling The longitudinal strains on Gleeble bar centerline are tensile with a maximum on eir side of central hot zone The transverse strains are compressive but vary with rmal treatment to a higher degree than variations in longitudinal strains The difference between strains at centerline and off centerline can be significantly greater than statistical error in aircooled Gleeble bars The strains in Gleeble bar have a high tensile component parallel to direction of maximum heat transfer (viz along bar axis) By contrast, large tensile strains in heataffected zone (HAZ) of weldment are along weld line which is essentially perpendicular to direction of maximum heat transfer The simulated conditions present in Gleeble bar test specimens are different from that observed in weld HA2 RESIDUAL STRESSES THAT DEVELOP IN THE heataffected zone (HAZ) are associated with heating and cooling of base metal in vicinity of fusion zone The rmal cycle experienced in HAZ can be simulated by Gleeble test The microstructures produced in Gleeble test bars are similar to microstructures found in HA2 of weldment, and it is of interest to show that mechanical properties in HAZ are equally well simulated in Gleeble specimens One of first residual stress investigations of Gleeble test bars was carried out with Xray residual stress analysis of austenitic stainless steel (1) The Xray method probes surface residual stress state, and material removal permits indepth characterization In this paper development of residual strains in 2 1/ Cr1 Mo steel Gleeble bars is examined with neutron strain scanning and related to strain mapping done in a 12cm (in) multipass weld made from same material (3) The 2 1/ Cr1 Mo steel has a bainitic microstructure with a bainite start temperature of approximately C The microstructures and hardness variations within a weld of a closely related steel (3Cr1MoOlV) have been described by Vitek and David (2) Generally, fusion zone consists of uniform bainitic microstructure with some retained austenite In HAZ re is a transition region consisting of a mixed bainite and ferrite structure with some retained austenite and n a return to uniform bainite having no retained austenite There is a hardness minimum in region containing ferrite, and location of this minimum depends on rate of cooling The formation of bainite upon cooling generates residual stress which is added to rmal expansion effects Residual stress contribution from austenite to bainite transformation is due to transformation plasticity associated with shear and dilatation of bainite transformation (3,) This paper presents a short description of neutron strain scanning method as applied to welded plate and Gleeble bars The contributions of both mechanical and chemical sources of lattice strain can play a part in study of weldments, thus measurements of "strainfree'' pieces taken from Gleeble bars are presented in this investigation as well The effects of solid state transformation are demonstrated by comparing femtic and austenitic plates Finally, strain maps in Gleeble bars are

2 "A interpreted, and strain map of welded bainitic steel plate is related to Gleeble bar results Neutron Scattering Experiments The use of neutron diffraction for residual stress evaluation is well established at most neutron research centers in world The method determines lattice strain from observation of changes in lattice dspacing obtained from diffraction peak shifts The advantage of neutron radiation is that samples can be probed to depths up to a few centimeters A full threedimensional strain tensor can be determined by appropriate variation of diffraction geometry However, measurements in a strainfree standard are needed for accurate residual stress results Strain mapping is based on using a small 1 The incident beam is defined by a shielding aperture and illuminates a lon column in scattering specimen The d&raction of neutrons in direction of lower arrow is n limited by a secon8 aperture, and reby a small diffracting volume is defined Fi diffracting region as shown schematically in Figure 1 The neutron strain scanning experiments were done at High Flux Isotope Reactor at Oak Ridge National Laboratory The measurements were done on a 127 mm (in) thick plate containing a multipass weld and three Gleeble bars with dimensions 127mm (in) x 127 mm (in) x 18 mm (in) The compositions of 2 1/ Cr1 Mo base metal and filler alloy are given in Table 1 The rmal treatment of Gleeble test bars is given in Table 2 Table 1 Commsition of base and filler metal, wt % Base Filler P1 P2 P3 C 11 9 Si 23 6 Mn P S Cr Mo 9 1 Ni 1 Fe Bal Bal Table 2 Thermal histories of Gleeble test bars (base metal) Peak temperature, O C Hold time, s Cooling Rate, OCImin

3 The neutron spectrometers used for this work were adapted for residual strain scanning by addition of beam collimators, a specimen translation stage, and a linear positionsensitive proportional counter() The (211) reflection from femte (bodycentered cubic) was used for dspacing determinations, and diffracting volume was typically 2 mm x 2 mm x 2 mm The M and P3 Gleeble bar samples were measured at HB3 spectrometer using a wavelength of 1 A The P1 and P2 Gleeble bar samples and an asreceived bar were measured at HB2 spectrometer using a wavelength of 16 8, Stressfree samples with dimensions 127 x 6 x 17 mm were cut from P1, P2, and P3 Gleeble test bars by electrodischarge machining These samples were measured at HB2 spectrometer using a wavelength of 161A Different from first two experiment sets, scattering volume in third experiment set was 1x 1x 1 mm Measurements of dspacing variations of se samples near center of hot zone in three Gleeble bars assessed variations in lattice parameter changes due to changes in chemical composition constituents However, contribution of chemical composition to dspacing variations in P1, P2, and P3 Gleeble bars appears to be negligible in Figure The largest variation is seen in samples taken from P1 which had most severe rmal treatment The observed strain variation shown in Figure is much smaller than range of total strains seen in Figure 3 The Gleeble bars P2 and P3 were investigated more thoroughly by measuring strains on section planes along length of bar In Gleeble bar M, strains on each of section planes were uniform in close agreement with value determined at center of section plane However, in Gleeble bar P3 which was air cooled, strains on section plane showed significant variations compared to value determined at center of plane Figure shows transverse and longitudinal strains as a function of distance along Gleeble bar The scatter in both transverse and longitudinal strains for any given plane is not symmetrical about strain measured at center It is apparent that strain system is complex and that strain variations from center to side within Gleeble bar are dependent on dwell time at 9 C and cooling rate Results of Strain Scanning Figure 2 shows strain scanning data comparing femtic and austenitic welds (3,) Longitudinal strains were measured at various distances normal to weld centerline averaged over several determinations at different depths in plate The peak longitudinal strains are comparable, but longitudinal and transverse (or radial) strain in ferritic weld exhibits a shoulder not seen in austenitic weld Figure 3 shows strain measurements made along centerline of Gleeble bars that were given rmal treatments described in Table 2 The most severe treatment in P1 which consisted of a s dwell time at 9 C followed by air cooling, results in high longitudinal tensile strains and small transverse strains In P2, which had a Imin dwell time at 9 C and cooling at 1 C per minute, this Gleeble bar shows smaller longitudinal tensile strains and a variation in transverse compressive strains In P3, with a 1min dwell time at 9 C followed by air cooling, this Gleeble bar shows higher tensile strains for longitudinal component, but not as large as in P1, and a sharp increase in transverse compressive strains Gleeble bars of 2 1/ Cr1 Mo alloy subjected to variations in rmal treatment are known to exhibit varying microstructures containing different phase 2 I ~~ Distance from Weld Center, mm Fig 2 The longitudinal strain component for ferritic weld (solid) shows a dip and a shoulder not seen in in multipass austenitic weld (dashed)

4 2 3 3 f 2 2 Y 7 =e E lo z ze Distance Along Bar, mm to + pi i, IIiii; a A " 3 2 T ' _ ; ; E zz 1 a 1 a A ; ; ; Distance Along Bar, mm Distance Along Bar mm 2 Fig 3 The longitudinal (filled) and transverse (open)strain components are shown for M ( s dwell time and air cooled), P2 (1min dwell time and 1CNIC/min coolin and P3 (lmin dwell time and air cooling The longitudinal strains show a tensile rise in heataffected zone and longitudinafcomponent in P3 goes sharply compressive The longitudinal strain peaks in Gleeble bars shown in Figure 3 are located away from center of hot zone in test bars The location of peaks undoubtedly is boundary between hot region undergoing bainite formation from austenite and cooler region in which initial bainite structure is only tempered The magnitude of se tensile longitudinal peaks and behavior of compressive transverse strain depend on both dwell time at 9 C and cooling rate It is tempting to assume that longitudinal strain variation observed in Gleeble bars can be applied directly to interpretation of longitudinal strains in weldment However, e i i differences in mechanical constraints, geometry, and iiii B 1 rmal conditions between weldment and + 2 i i i i i Gleeble bar appear to influence orientation of + strain tensor The longitudinal tensile strain in E 2 Gleeble bar is parallel to direction of heat f flow along length of bar, while tensile I 2 i "+ longitudinal strain in HAZ is perpendicular to inii i"i"ini~ flow, which is nto base direction of heat metal The difference in tensile strain orientation between weld and Gleeble bar undoubtedly arises Distance from center, mm from absence of transverse mechanical constraint Fig The variation in "stressfree" dspacings is in Gleeble bar The bainite reaction can be represented as a strain The variations for all three Gleeble expected to choose quite different transformation samples are comparable to estimated error in shear variants in response to different determination of strains in test bars and thus indicate that chemically induced strain is small compared to mechanical constraints in two cases ~ j r mechanical strain of residual stress transverse and longitudinal strains for any Oven plane is not symmetrical about strain measured at t% center

5 a 8 6 c r i l f i 6 bi;,aii i; ' " " o Distance, mm 2 Y = o f w Distance, mm Fig The transverse (top) and longitudinal (bottom) strain components measured on Gleeble bar centerline define continuous line drawn in both figures The points in groups centered away from line are result of measurements made at locations displaced from bar center axis summary Fastcooled Gleeble bar (P3)has a complex strain pattern which indicates influence of lateral heat losses that are not seen in slowcooled (P2) Gleeble bar Short dwell time appears to sharpen strain variations The development of a tensile peak in longitudinal strains in Gleeble bar suggests an explanation for "broadening" of residual stress pattern in femtic relative to austenitic strain pattern Even though Gleeble test is known to produce same microstructures found in welds of same material, we have shown a difference in orientation of strain tensor in weld and in Gleeble bars The interpretation of properties in Gleeble bars, which are sensitive to residual stress, must be done with proper consideration of se differences Acknowledgments This research was sponsored by US De artment of Energy, Assistant Secret for Energy Efcien and Renewable Ener Office?Transportatatlon T'echnxo 'es, as part of HigTemperature Materials Laboratory 8ser Program; by Assistant Secretary for Energy Research, Office of Basic Energy Sciences, Division of Materials Sciences for sup ort for High Flux Isotope Reactor facilities; and by tke Laboratory Directed Research and Develo ment Program The very able sup rt in pre aring is wepded lates and Gleeble specimens CRobb Lowledged The authors also thank Spence g References (1) V S Iyer, R W Hendricks, and S A David, Advances in XRay Analysis, Vol 3, ed C S Barrett et al, Plenum Press, New York, 1991; p 623 (2) M Vitek and S A David, Metall Trans A, 21A, (199) (3) H Root, T M Holden, Schroeder, C R Hubbard, SSpooner, T A Dodson, and S A David, Mater Sciand Technol 9, 79 (1993) () S Spooner, X L Wang, C R Hubbard, and S A David, pp in Proceedings of th Zntlernational Conference on Residual Stresses, (Baltimore, Maryland, une 81?1991, Society for Experimental Mechanics, Inc, Bel, CT 199 () H K D H Bhadeshia, S A David, M Vitek, and R W Reed, Materi Scie and Technolo 7, ((1991) DISCLAIMER This report was prepared as an account of work sponsored by an agency of United States Government Neir United States Government nor any agency reof, nor any of ir employees, makes any warranty, express or implied, or assumes any legal liability or respnsibility for accuracy, completeness, or usefulness of any information, apparatus, product, or process disclosed, or represents that its use would not infringe privately owned rights Reference herein to any specific commercial product, process, or service by trade name, trademark, manufacturer, or orwise does not necessarily constitute or imply its endorsement, recom mendation, or favoring by United States Government or any agency reof The views and opinions of authors expressed herein do not necessarily state or reflect those of United States Government or any agency reof