HAZ Toughness of Ti-Microalloyed Offshore Steel in As-Welded and Simulated Condition HAZ žilavost mikrolegiranih Ti-ofshore jekel v varjenem in simuliranem stanju Rak I.1, Faculty of Mechanical Engineering, Maribor M. Kogak, B. Petrovski, GKSS Research Center, Geesthacht V. Gliha, Technology Research Center, Maribor The CTOD values measured on 34 mm thick SENB specimen taken from multipass 1/2K joints were compared with the values obtained from 8 mm thick SENB specimens with simulated microstructures of CGHAZ (a/W=0.5). Single and double thermal cycles were used to simulate various thermal treatment vvhich HAZ at the weld bond may experience during the welding. The CTOD fracture toughness testing of the simulated specimens can provide toughness values not affected by the mechanical heterogeneity (strength mis-match betvveen weld and base metals) provided one can simulate the microstructure of interest correctly. The examinations of these simulated specimens show the presence of the local brittle zones (LBZ) in spite the steel was Ti-microalloyed. An attempt to correlate CTOD and Charpy impact toughness values on simulated microstructures was undertaken. Key words: HAZ toughness, Ti-microalloyed steels, fracture mechanics, weldability Introduction It is well known that the coarse grained heat affected zone-CGHAZ region of many structural steel vvelds can be the least tough region of the weld joint. In the literature survey\ a huge attempt can be no-ticed for improving the CGHAZ toughness of the modern microalloyed steels. The steel makers suc-ceed to reduce the coarse grain size and the vvidth of CGHAZ and hence to improve the toughness proper-ties by using Ti as microalloying element two mecha-nisms mainly1'2'3'4: - Grain boundary pinning by uniformly distributed TiN particles, sized from 0.02 um to 0.08 jum, but keeping the Nb and V contents low. Presence of fur-ther alloying elements serving as nitride formers gen-erally tends to decrease the stability of TiN particles and increases the particle size. - Grain boundary pinning by uniformly distributed stable Ti203, sized from 0.5 um to 3 um but promoting also interferential nucleation of acicular ferrite as main beneficial effect. Although TiN is thought to be comparatively stable even at high peak temperatures, complete and/or partial dissolution (depending on the size and com-position of the precipitate) can stili be expected, Dr. Inoslav RAK, dipl.inž. Faculty of Mechanical Engineering Maribor, Smetanova 17 since, TiN particles can occur in various sizes rang-ing from several microns to several hundred angstroms. Hovvever, a particle size smaller than 0.1 |im has been found to be the most effective in grain boundary pinning. Therefore, TiN can only be effective in suppressing grain coarsening in the HAZ if the method of vvelding, the ratio of Ti/N and the level or presence of other microalloying elements produce Ti precipitates of appropriate size and distribution. Ti-oxides are efficiently used in improving the toughness of steel and also in weld metal deposits due to their dual role, restricting the grain boundary migration and acting as nucleus for acicular ferrite formation (since they are more stable than TiN precipitates at higher temperatures). A sufficient amount of precipitates should also remain undissolved in the HAZ and these should act as pinning and ferrite nucleation sites. Hence, optimum numbered and sized fine TiN or Ti-oxide precipitates must be present if an improvement on HAZ microstructure/toughness of the zone adja-cent to weld metal is expected to occur. In the present paper, the measurement of HAZ fracture toughness of a SAW joint of 40 mm thick TiN treated offshore steel was undertaken. Fracture mechanics CTOD tests were conducted on 34 mm thick specimens taken from multipass 1/2K joint and the values were compared with the values measured on 8 mm thick SENB specimens containing different simulated HAZ microstructures. The aims of this work were: - to measure the toughness of different HAZ mi-crostructures by using simulated microstructures by omitting the problem of crack/notch tip location in the specimens taken from actual joints, - to compare the CTOD results of simulated specimens with CTOD values at stable crack initiation, (CTODi) values, obtained from full thickness specimens taken from a SAW joint vvhich was about 27% overmatching, - to correlate the CTOD critical values 8c and the CTOD values at the onset of stable crack grovvth 8i with Charpy test values, - to discuss the possible effect of weld metal strength overmatching of a real weld joint on frac-ture behavior and on fracture toughness values. Experimental details Parent Material and VVelding Procedure The C-Mn steel used was in normalized condition and its chemical composition is given in Table 1 vvhich gives also the ratios of Ti/N and C/N. The steel has low C, S and is alloyed with Ni. The mechanical properties of the 40 mm thick steel and SA weld metal are given in Table 2. The plates vvere vvelded using Tandem SAW multipass procedure vvith a single bevel butt vveld preparation as shovvn in Fig. 1. The vvelding procedure is given in Table 3. Table 1: Chemical Composition of the Steel StE 355Ti in % Steel Type C Si Mn P S N Al 0.09 0.43 1.46 0.013 0.003 0.0071 0.046 StE 355+Ti Cu Ni Ti Nb Ca C/N Ti/N 0.12 0.44 0.016 0.022 0.001 12 2.25 Pcm=0.190, Ceq=0.370 Table 2: Mechanical Properties of the Parent Steel and SAW Weld Metal Steel Type (MPa) (MPa) (%) Charpy V impact energy (J)' -10°C -40°C Mismatch factor M StE 355 Ti 388 515 Weld metal 492 578 32.5 24.4 292 300 166 140 1.27 "Transverse direction Table 3: SAW VVelding Procedure Tandem-SAW vvelding procedure Number of passes 10 Wire/flux OE-SD3/OP121TT Heat input 40 kJ/cm, At8/5=40 sec Preheating temperature 100°C Interpass temperature_200° C_ Figure 1: SA vveld cross section Thermal Simulation Procedure The specimens for microstructural simulation of the HAZ region vvere extracted from the parent plate in the rolling direction vvith the dimensions of 8x15x70 mm. Various single and double thermal cycles vvere carried out to simulate different HAZ microstructures. The peak temperatures of the simulations are given in Table 4. Single cooling tirne (At8/5=40 sec) was used for ali thermal cycles vvhich corresponds to the SAW condition. HAZ Fracture Toughness Testing The HAZ toughness of the multipass vveld and simulated microstructures was measured using Charpy V-notch and CTOD specimens. The single edge notched bend (SENB) CTOD specimens vvere ma-chined from as-vvelded multipass joints in Bx2B geometry (B=34 mm) notched in through-thickness in the HAZ and tested at -10°C. For the simulated microstructures, the SENB specimens vvere 8 mm thick and vvere approx. Bx2B type. During the CTOD tests the DC potential drop technique vvas used for moni-toring the stable crack grovvth5. Load line displacement (VLL) vvas also measured vvith a reference bar to minimize the effects of possible indentations of the rollers. Fatigue precracking vvas carried out vvith "step wise high R-ratio" (SHR) procedure for ali specimens6'7. This technique is suc-cessfully used at GKSS Research Center for as vvelded specimens to obtain a uniformly shaped fatigue precrack. The SHR technique uses two stress ratios, R=0.1 for crack initiation and grovvth of about 1 mm then, stress ratio of R=0.7 vvith the allovvable maxi-mum load, until the required a/VV ratio is obtained. The CTOD values vvere calculated in accordance vvith BS 5762 (5BS)8 and in the čase of real vveld specimens also directly measured vvith GKSS developed 55 clip gauge on the specimen's side surface at the fatigue crack tip over gage length of 5 mm9. Table 4: HAZ Simulation Procedure Data for Single and Double Thermal Cycles Specimen First cycle At8/5 Second cycle At8/5 designation Tp1 (°C) (S) Tp2 (°C) (s) 0 1380 40.2 _ _ A 1370 40.0 705 82* B 1390 40.5 907 43 C 1380 40.0 960 42 D 1380 40.0 1025 41.5 E 1360 40.0 795 80+ * instead At8/5, cooling tirne At6/3 was measured; At8/5=1/2At5/3 Results CTOD Results The CTOD results8 obtained from SENB speci-mens extracted from multipass welds and simulated microstructures are given in Fig. 2. The critical values of CTOD and CTOD data for the initiation of the duc-tile tearing (8i) are shovvn in Fig. 3 (see also the ex-planation of symbols in the tables 4 and 6). The 8i value is defined as the value of CTOD for the crack grovvth of 0.2 mm in accordance with the ESIS procedure10. In these figures, the CTOD values of the HAZ multipass weld joint (F) can be compared with the results of the simulated microstructures (O-E). 3- m iS 2 i 0 6 -H c : ♦ □□8n Figure 2: HAZ "apparent" fracture toughness of simulated and SAW microstructures 0.8. 0.7. 0.6 0.5 E 04 E S 0.3 O u 0.2 0.1 ( Steel StE 355+Ti ) A V qP A ocbo i i-i-T-i- O A B C D *Fracture before initiation al aa=0.2 mm - (8c) Figure 3: HAZ "intrinsic" fracture toughness of simulated and SAW microstructures After CTOD testing, the post-test sectioning and mi-crostructural examinations were conducted for ali specimens to identify the fatigue crack tip microstruc-ture and the location of the initiation. Simulated and Real Multipass HAZ Microstructures Impact Toughness Testing The Charpy impact toughness for the simulated microstructures are presented in Table 5a. Fracture appearance transition temperature-F ATT and the maximum hardness values obtained for each mi-crostructure are included in Table 5b. The same data obtained from multipass weld HAZ are presented in Table 6. Table 5a: Charpy Toughness of Simulated Microstructures Specimen Energy (J)' designation -60 -40 -20 0 20 40 60 0 - 13 20 68 176 _ _ D 17 38 128 - 210 253 - C 18 23 85 - 186 220 - B 22 33 119 - 197 - - E - 11 32 - 126 152 - A - 13 13 - 55 - 135 'average of three specimens Table 5b: FATT, Hardness, Shift Temperature, CTOD Transition Temperature and Calculated Critical CTOD Value Specimen FATT Hardness AT FATT-AT 8c designation (°C) HV10 (X) ro (mm) 0 +11 213 31 -20 0.13 D -8 210 38 -46 0.15 C 0 204 41 -41 0.14 B -6 208 39 -45 0.15 E +29 225 25 + 4 0.14 A +43 230 24 +19 0.10 Table 6: SA VVeld Joint HAZ Charpy Impact Toughness and Hardness Values Specimen Location Energy(J)" Energy(J)' Hardness designation -10°C -40° C HV10 Cuplayers 197 149 189 F Middle layers 191 156 - Root layers 179 103 179 ' average of three specimens Discussion Charpy-V Test Results It is clear that high Charpy-V impact toughness values of the real HAZ at the weld bond (Table 6) are the average toughness of several microstructural regions due to the relatively large machined notch tip radius vvhere more ductile areas of HAZ also contribute to the deformation and fracture. This implies that the Charpy-V test produces unreliable results if quantita-tive CGHAZ/LBZ toughness of various multipass vvelds is going to be assessed. This can be proved by Charpy impact toughness values obtained from sim-ulated specimens with uniform microstructures (Table 5a). Different HAZ regions represented by various thermal simulation procedures exhibit different toughness and hence varying sensitivity for LBZs appearance at the testing temperature. It has to be pointed out that the cause for difference of real multipass and simulated HAZ impact toughness can not be the deviation of cooling tirne. The analyzed de-pendance of impact toughness and cooling tirne shovvs only a slight change in the range of 30 - 50 sec. CTOD Test Results The standard CTOD fracture toughness results (Fig. 2) show much higher toughness values (F) for the specimens extracted from multipass weld joint if they are compared with the values of simulated specimens. It was expected to obtain similar or even some better toughness values by conducting measurement on the six different types of simulated microstructures. But the measurement on the thick SENB specimens of multipass weld joints did not show any low CTOD values. On the other hand, comparison of the "intrinsic" fracture toughness values (crack initiation at Aa=0.2 mm) obtained on both specimen types is better (Fig. 3 - B, C and D), due to its size indepen-dent nature. But even in this čase, the fracture toughness of real SA weld joint is slightly higher than in ali simulated cases. The reason is the full sampling of the CGHAZ microstructural constituents in simulated specimens compared to the full thickness CTOD specimens extracted from real vveld joints, despite of higher constraint and overmatching effect in the lat-ter, which should lovver the CGHAZ fracture toughness. The lovvest fracture toughness values were es-tablished by the simulated unaffected coarse grained (UCG) HAZ - single cycle microstructure designated by O. The second thermal cycle applied betvveen AC1 and AC3, (E) and belovv AC1, (A), did not im-prove the toughness of the UCGHAZ. The second thermal cycles above AC3, (B, C and D), as expect-ed, have improved the fracture toughness due to the refinement of the UCGHAZ, but the fracture toughness level of the real vveld HAZ was stili not reached. It can be concluded that the HAZ fracture toughness measured is highly effected by crack tip microstructure. The lovvest fracture toughness can be obtained by positioning the crack mainly into the CGHAZ microstructure. If this čase is combined vvith the highest constraint condition the cleavage crack initiation can occur from the CGHAZ of the real vveld joints by dislocation piling up mechanism suggested by the RKR local fracture criterion11'12. It is evident that the estimation of "intrinsic" fracture toughness on simulated specimens for different HAZ regions can be rather informative to control the LBZ susceptibility of the steel even if the CTOD results do not indicate any embrittlement in the multipass vveldments. The fracture behavior of the LBZs can be influ-enced by the strength mismatching of the vveld joint. High strength and tough vveld metal can provide a possibility for the HAZ notched CTOD specimens to fracture in unstable fashion. If the fatigue crack tip is located in the vicinity of the CGHAZ (in the over-matched vveld metal having good toughness), the brittle crack can stili initiate at the CGHAZ under the influence of the strength mismatching, as shovvn in Fig. 4. Higher strength vveld metal side of the specimen vvill not allovv the plastic zone at the crack tip to develop, because of softer base material and the consequence is a constraint raising at the CGHAZ. Consequently, the CGHAZ reaches the critical condition at lovver nominal stress/load and therefore fracture may predominantly initiate and remain in the brittle CGHAZ as shovvn in Fig. 4. Fracture behavior of the CGHAZ should also be examined in terms of me-chanical heterogeneity of the vveld joint since, identi-cal CGHAZ microstructure can give substantially different toughness values (i.e. apparent) if one varies the vveld metal strength mismatch. Figure 4: Fatigue crack tip in the overmatched vveld metal but brittle fracture initiates and remains at CGHAZ during the CTOD test CTOD Fracture Toughness and Charpy tmpact Toughness Correlation In general, for medium strength steels and medi-um thicknesses, the Charpy transition curve moves relatively to the higher temperature compared to that of the CTOD. The Charpy transition temperature is defined by F ATT. In the čase of CTOD, the transition temperature is assumed to be the temperature at which the critical CTOD value (8c) becomes equal to the critical value for the onset of stable crack grovvth (8i). The difference of these two transition tempera-tures is sensitive to the strain rates and noch acuity of the impact Charpy and CTOD tests and the yield stress of the material and the thickness of the spec-imen respectively13. AT = 133 — 0.125