- Email: [email protected]
Effect of microstructure on mechanical properties in weld-repaired high strength low alloy steel
Materials and Design 36 (2012) 233–242
Contents lists available at SciVerse ScienceDirect
Materials and Design journal homepage: www.elsevier.com/locate/matdes
Effect of microstructure on mechanical properties in weld-repaired high strength low alloy steel Chunguo Zhang a,b,1, Xuding Song a, Pengmin Lu a,⇑,1, Xiaozhi Hu b a b
School of Construction Machinery, Chang’an University, Xi’an 710064, PR China School of Mechanical and Chemical Engineering, The University of Western Australia, Perth 6009, Australia
a r t i c l e
i n f o
Article history: Received 20 September 2011 Accepted 8 November 2011 Available online 15 November 2011 Keywords: D. Welding F. Microstructure G. Fractography
a b s t r a c t To understand the effect of microstructure on mechanical properties of weld-repaired high strength low alloy (HSLA), as-received and weld-repaired HSLA with and without buffer layers (BLs) were prepared. Microstructure analysis was carried out using optical microscope and SEM, and mechanical properties were measured by Vickers hardness test and fatigue test. The fatigue resistance of weld-repaired HSLA without BL was deteriorated with comparing to parent metal (PM). Meanwhile, Vickers hardness (VH) showed an obviously reduction in the melted parent metal (MPM), which was due to formation of predominately block ferrite. For the weld-repaired HSLA with BL, the VH and fatigue resistance increased with the incorporation of 4 mm BL, which was mainly due to formation of lath ferrite and fine-grained pearlite and bainite. When BL thickness increased to 10 mm, the VH and fatigue resistance decreased, which was because the thick BL diluted the MPM. VH number from low temperature (below melting point) heat affected zone (HAZ) fluctuated, but had a little scatter. However, the fatigue crack growth rate from HAZ was not obviously affected by the welding as comparison with the PM. Ó 2011 Elsevier Ltd. All rights reserved.
1. Introduction High strength low alloy (HSLA) steels with excellent strength characteristics have been widely used in various applications including cars, trucks and cable-stayed bridges. The strength of HSLA steels comes from their microstructure. Strength is increased by: increasing the amount of pearlite and bainite, increasing the fineness of the grains structure, increasing the amount of hard precipitate [1]. Wear damaged sections of supporting structures manufactured from HSLA are often weld-repaired or filled by similar metals through welding [2], or weld deposited by hard-facing alloys to provide a wear resistant surface for structural base materials. However, HSLA steels are hard to weld repair and special care is required because of their high strength [3,4]. Furthermore, the welding joints of HSLAs are often become the weak link in mechanical testing due to softening in the welding joints. To understand the detrimental influence of the welding on mechanical properties of welded HSLAs, a number of studies have been performed to
⇑ Corresponding author. Address: School of Construction Machinery, Chang’an University, Nan Er Huan Zhong Duan, Xi’an 710064, PR China. E-mail addresses: [email protected] (C. Zhang), [email protected] (X. Song), [email protected] (P. Lu), [email protected] (X. Hu). 1 Tel./fax: +86 29 82334322. 0261-3069/$ - see front matter Ó 2011 Elsevier Ltd. All rights reserved. doi:10.1016/j.matdes.2011.11.016
assess the effects of chemical compositions, metallographic structure, welding procedure and heat treatment (pre- and post-). With increasing of the copper content in weld metal, microstructure became finer in all zones of the weld metal, hardness and tensile strength increased, but Charpy impact toughness decreased [5]. The increase of manganese content was useful to refine and homogenize the weld microstructure, and the increase of titanium content in the range of 0.02–0.08% increased the acicular ferrite for the API 5L-X70 (HSLA) with moderate manganese content [6]. The amount of M-A constituent in coarse-grained HAZ of HSLA influenced the toughness value, and voids and micro-crack initiated at the M-A constituents [7]. The high-toughness low-carbon bainite in the HAZ of EH36 TMCP decreased fatigue crack growth rate, and post weld heat treatment was unnecessary [8]. Due to the coarse upper bainite, the weld metal was the weakest link in terms of toughness and fatigue resistance [9]. The fusion zone of laser welded HSLA, near to the interface of fusion zone and HAZ, was the weakest link of fatigue behavior [10]. Under the condition of gas metal arc welding, the high hardness welds (over 248HV) of HSLAs (API 5L X70 and X80) increased the resistance to sulfide stress cracking at low H2S concentrations [11]. Welded HSLA is often prone to softening in HAZ, which causes a weak link in mechanical properties [12]. However, HAZ of HSLA showed lesser softening in low heat input welding than in high heat input welding, and post weld heat treatment about AC3 and
234
C. Zhang et al. / Materials and Design 36 (2012) 233–242
external cooling were beneficial to reduce the tendency for softening in HAZ [13]. Reducing cooling time decreased the austenite grain size and volume fraction of M/A constituent, thus increased the toughness of HAZ of HSLA [14]. The cooling rate changed the microstructure in weld metal was also reported by Ghosh, et al. [15]. In the coarse-grained region of HAZ, the microstructure consisted of predominantly bainite packets and a small proportion of acicular ferrite was more stable than bainite during tempering [16]. Submerged arc welded joint of HSLA (14HNMBCu grade S690Q) had higher resistance to hydrogen degradation in sea water environment than shielded metal arc welded joint [17]. With different emphases to those aforementioned studies on welded HSLA steels, buffer layer between weld metal and parent metal was introduced to increase the fatigue resistance of welding joint in this paper. The primary objective of the present study was to investigate the metallographic structure of extensively weldrepaired HSLA steel with or without a buffer layer, and its influence on the mechanical properties of welding joint, which has not been systematically studied as shown in previous literatures.
imens’ surfaces after polishing and etching were obtained using professional digital camera. The micro-structure of fatigue surfaces of tested specimens associated with various regions in as-received and welded specimens, especially in MPM and BL + MPM, were observed in detail by a scanning electron microscopy (SEM).
Table 2 Mechanical properties of PM, WM and BL. Series
Tensile properties Yield strength (Mpa)
Tensile strength (Mpa)
Elongation (%)
Impact energy (J)
PM
690 min
790 min
18 min
WM
690 min
760 min
17 min
BL
410 min
490 min
22 min
40 min ( 20 °C) 50 min av. ( 51 °C) 27 min av. ( 29 °C)
2. Experimental procedure 2.1. Material and specimen The substrate material used for the present investigation was a HSLA steel (Bisplate80-parent metal PM) with a low carbon content. The weld metal (WM) and buffer layer (BL) were SmoothCor™ 115 and SmoothCor™ 70C6, respectively, thus effectively forming a welding joint together with PM with different strength characteristics. Flux cored arc welding was employed to perform the extensive welding repair, while 100% CO2 was used as the shielding gas. The core wire had a diameter of 1.2 mm, and electrode stick-out was 20 mm. A constant current power source (230A, 27 V) with DC + polarity was used in all cases (WM, 4 and 10 mm BLs). The chemical compositions and mechanical properties of PM, WM and BL are listed in Tables 1 and 2. Extended-Compact Tension (E-CT) specimens with machined through-the-thickness notches were machined following the specifications of ASTM E647 (23) [18]. The specimens were divided into four groups: (1) as-received HSLA, (2) weld-repaired HSLA without BL, (3) weld-repaired HSLA with 4 mm BL, and (4) weld-repaired HSLA with 10 mm BL. Dimensions of the E-CT specimens are shown in Fig. 1, and all dimensions are in mm.
(b) Weld-repaired HSLA without BL
(a) As-received HSLA (mm)
2.2. Microstructure study The four groups of E-CT specimens, (1) as-received HSLA specimens, (2) weld-repaired specimens without BL between the WM and PM, (3) weld-repaired specimens with 4 mm BL, and (4) weld-repaired specimens with 10 mm BL, were mechanically polished and etched using 4% nital solution for a few seconds to reveal their microstructures. The microstructures of the four groups of E-CT specimens were examined in detail by inverted optical microscope (JVC CMM-22E). The macrographs of the weld-repaired spec-
(d) Weld-repaired HSLA with 10 mm BL
(c) Weld-repaired HSLA with 4 mm BL
Fig. 1. Dimensions of E-CT specimens.
Table 1 Chemical compositions of PM, WM and BL. Series
PM WM BL
Element (wt.%) C
Si
Mn
Cr
Mo
Ni
P
S
B
CE(IIW)
0.18 0.06 0.03
0.20 0.30 0.59
1.40 1.40 1.66
0.20 0.22 –
0.20 0.44 –
– 2.29 –
0.01 – –
0.003 – –
0.0010 – –
0.50 0.58 0.31
CE(IIW) = C + Mn/6 + (Cu + Ni)/15 + (Cr + Mo + V)/5.
C. Zhang et al. / Materials and Design 36 (2012) 233–242
2.3. Mechanical properties The Vickers hardness variation form the WM, BL, melted parent metal (MPM) during welding, HAZ to PM in weld-repaired HSLA was measured with an indenting load of 20 kg by using a hardness tester (Mitutoyo AVK-C2). The four groups of E-CT specimens were tested under the same fatigue loading condition using Instron 8501. Constant amplitude loading with haversine waveform at a frequency of 5 Hz was used, and the R ratio was set at 0 throughout the test. The crack length was measured from the loading line as indicated in Fig. 1b. The Vickers hardness profiles and Paris fatigue curves for each group were measured twice with two identical specimens, and the results were quite consistent. 3. Results Microstructure of as-received and weld-repaired HSLA was examined using optical microscope with special emphasis on the regions in WM, MPM, HAZ, and around the welded interfaces and fusion line (FL). To correlate the microscopic observations to the relevant locations and corresponding VH and da/dN measurements, sketches for the boundary regions, VH distribution curves and da/dN curves were numbered accordingly for the weldrepaired HSLA with and without BLs. 3.1. As-received HSLA Following the above descriptions, Fig. 2 displays the details for PM. The VH distribution curves are shown in Fig. 2a, the results are quite consistent ranging from 224VH to 264VH, which can be as a good reference to study hardness variation in the weld-repaired HSLA with and without BLs. Paris fatigue curves from the as-
235
received HSLA are shown in Fig. 2b. Fig. 2c is an optical micrograph of the PM, showing a microstructure consisted of ferrite, bainite and pearlite. The VH measurements and fatigue propagation tests from two separate specimens were very consistent, which therefore provided a good reference to study effects of welding, incorporation of BL and various BL thicknesses. 3.2. Weld-repaired HSLA without BL Fig. 3 displays the macrograph, VH distribution curves, Paris fatigue curves, relevant locations and microstructure details for the weld-repaired HSLA without BL. The macrograph of surface after etching is shown in Fig. 3a, which displays an obviously white band of approximately 5 mm in width. The band was the melted parent metal (MPM) during welding. This was due to the carbon loss because of the over high temperature during the welding. As shown in Fig. 3f, the micrograph of MPM (white band) consisted of predominately block ferrite and some pearlite. The VH number in the corresponding position decreased. This is because carbon and alloy elements in MPM changed during welding. The microstructure of predominately block ferrite produced obvious reduction in VH shown in Fig. 3b, and greatly increase of fatigue crack propagation rate (da/dN) shown in Fig. 3c. Microstructures of distinct regions (WM, MPM and HAZ) in the weld-repaired HSLA without BL are shown in Fig. 3d–i. The microstructures of the WM in a multi-pass weld were dependent on the homogeneity of composition and thermal history of welding. The as-welded microstructures were reheated by the subsequent passes, thus different structures were obtained according the related thermal cycles and chemical compositions. The microstructure of the WM far from the MPM was consisted of predominately pearlite, and some ferrite (Fig. 3d designated by ‘‘1’’). As shown in
(a) VH distribution curves
(c) Microstructure Fig. 2. As-received HSLA.
(b) Paris fatigue curves
236
C. Zhang et al. / Materials and Design 36 (2012) 233–242
Fig. 3b and c (designated by ‘‘1’’), the fatigue resistance and VH number were increased by increasing the amount of pearlite. In the WM near the MPM, the amount of pearlite decreased with increasing the amount of ferrite (Fig. 3e designated by ‘‘2’’). The
microstructure reduced the VH number and fatigue strength shown in Fig. 3b and c (designated by ‘‘2’’), respectively. The microstructures of the HAZ were dependent on the temperature during welding. As shown in Fig. 3b, VH number fluctuated in
(a) Macrograph
(b) VH distribution curves
(c) Paris fatigue curves
(d) WM far from the MPM
(e) WM near to the MPM
(f) MPM
(g) Overheated zone in the HAZ
(h) Normalized zone in the HAZ
(i) Partial phase transformation zone in the HAZ
Fig. 3. Weld-repaired HSLA without BL.
C. Zhang et al. / Materials and Design 36 (2012) 233–242
the HAZ because of the different structure, and was reduced in overheated and partial phase transformation zones. The micro-
structure of overheated zone was predominately block ferrite and some bainite shown in Fig. 3g. The block ferrite was harmful to
(a) Macrograph
(b) VH distribution curves
237
(c) Paris fatigue curves
(d) WM
(e) BL+MPM
(f) Overheated zone in HAZ
(g) Normalized zone in HAZ
(h) Partial phase transformation zone in HAZ Fig. 4. Weld-repaired HSLA with 4 mm BL.
238
C. Zhang et al. / Materials and Design 36 (2012) 233–242
fatigue resistance. But the da/dN was similar to the PM shown in Fig. 3c, this is because the stress gradient reduced the da/dN as the crack propagated from soft MPM to strong HAZ. In the normalized zone, a significant grain refinement (ferrite + pearlite) was observed (Fig. 3h), which increased the VH number. The
microstructure of partial phase transformation zone consisted of coarse-grained ferrite together with some pearlite (Fig. 3i), which reduced the VH number. As shown in Fig. 3c, the da/dN from HAZ was similar to the PM.
(a) Macrograph
(b) VH distribution curves
(d) WM far from the BL+MPM
(c) Paris fatigue curves
(e) WM near the BL+MPM
(f) BL + MPM
(g) Overheated zone in the HAZ
(h) Normalized zone in the HAZ
(i) Partial phase transformation zone in the HAZ
Fig. 5. Weld-repaired HSLA with 10 mm BL.
C. Zhang et al. / Materials and Design 36 (2012) 233–242
3.3. Weld-repaired HSLA with 4 mm BL Fig. 4 displays the macrograph, VH distribution curves, Paris fatigue curves, relevant locations and microstructure details for the weld-repaired HSLA with 4 mm BL.
(a) The interface of WM and MPM
(c) The interface of MPM and HAZ
The macrograph of surface after etching is shown in Fig. 4a, which displays an obviously dark band of approximately 4.5 mm in width. The band was the mixture of BL and MPM during welding, which was called BL + MPM. Because both the BL and MPM were thin, the mixture was homogeneous. The microstructure of
(b) Enlarged view of rectangle in (a)
(d) Columnar crystal (rectangle in (c))
Fig. 6. Weld-repaired HSLA without BL.
(a) The interface of WM and BL+MPM
(c) The interface of BL+MPM and HAZ
239
(d) Columnar crystal (rectangle in (c))
Fig. 7. Weld-repaired HSLA with 4 mm BL.
240
C. Zhang et al. / Materials and Design 36 (2012) 233–242
(a) The interface of WM and BL+MPM
(c) The interface of BL+MPM and HAZ
(b) Enlarged view of rectangle in (a)
(d) Columnar crystal (rectangle in (c))
Fig. 8. Weld-repaired HSLA with 10 mm BL.
Fig. 9. Comparison of the VH distribution curves of weld-repaired HSLA with and without BL.
the BL + MPM was ferrite (predominately lath ferrite together with some granular and small block), fine pearlite and some bainite, shown in Fig. 4e. As shown in Fig. 4b and c, the predominately lath ferrite together with pearlite and bainite increased both VH number and fatigue resistance. In the WM and BL + MPM, the mixed microstructures composed of lath ferrite, granular ferrite and some block ferrite together with pearlite and bainite (Fig. 4d and e) were existed. Other optical metallographic results (not shown here) indicated there was no significant micro-structural change between the WM and BL + MPM. The microstructure is beneficial to increase of VH and fatigue behavior, which was proved by showing high VH number (Fig. 4b) and low da/dN (Fig. 4c). Microstructures of distinct regions in the HAZ are shown in Fig. 4f–h. The microstructure of the overheated and normalized zones was consisted of fine-grained ferrite and pearlite. The microstructure of the partial phase transformation was consisted of block ferrite and pearlite. As shown in Fig. 4b, VH number in the HAZ fluctuated, and there was reduction in overheated and partial phase transformation zones. The da/dN in the normalized and partial phase transformation zones was similar to the PM, but slightly higher in overheated zone than that of PM. This was because the stress gradient accelerated the fatigue crack growth as the crack propagated from strong BL + MPM to relative soft HAZ. 3.4. Weld-repaired HSLA with 10 mm BL
Fig. 10. Comparison of the Paris fatigue curves of weld-repaired HSLA with and without BL.
Fig. 5 displays the macrograph, VH distribution curves, Paris fatigue curves, relevant locations and microstructure details for the weld-repaired HSLA with 10 mm BL. The macrograph of surface after etching is shown in Fig. 5a, which displays an obviously white band of approximately 8 mm in width. The band was the mixture of BL and MPM during welding, which was called BL + MPM. The microstructure of the BL + MPM was predominately block ferrite together with a little pearlite, shown in Fig. 4f. As shown in Fig. 5b and c, the predominately block ferrite reduced VH number and fatigue resistance.
C. Zhang et al. / Materials and Design 36 (2012) 233–242
The microstructure of the WM far from the BL + MPM was consisted of predominately pearlite, and some ferrite (Fig. 5d designated by ‘‘1’’). As shown in Fig. 3b and c (designated by ‘‘1’’), the fatigue resistance and VH number were increased by increasing the amount of pearlite. In the WM near the MPM, the amount of pearlite decreased with increasing the amount of ferrite (Fig. 5e designated by ‘‘2’’). The microstructure reduced the VH number and fatigue strength shown in Fig. 5b and c (designated by ‘‘2’’), respectively. Microstructures of distinct regions in the HAZ are shown in Fig. 5g–i. The microstructure of the overheated zone was predominately widmanstatten structure. This was due to repeatedly heating during welding. The widmanstatten structure reduced the strength, and increased the brittleness. This was proved by showing high VH number (Fig. 5b) and high da/dN (Fig. 5c). As shown in Fig. 5h, the grain size of fine-grained normalized zone (pearlite + ferrite) was significantly smaller than that of PM. In the partial phase transformation zone, the microstructure was consisted of ferrite, and pearlite. The different metallographic structures of distinct regions in the HAZ changed the VH number, but did not influence the fatigue resistance obviously.
4. Discussions
241
The grain size around welded interfaces in weld-repaired HSLA with 4 mm BL (Fig. 7) was significantly smaller by showing lath ferrite structure than that in weld-repaired HSLA without BL (Fig. 6) because of the incorporation of 4 mm BL. In the case of 10 mm BL, grain refinement was not obviously (Fig. 8). The lath ferrite structure was beneficial to increase of fatigue resistance and VH number (Fig. 4b and c).
4.2. Analysis on Vickers hardness Because the VH distribution curves in each group were quite consistent, one specimen in each group was selected to compare. As shown in Fig. 9, VH number from the weld-repaired HSLA with and without BL, fluctuated in the WM, MPM (BL + MPM) and HAZ, and had a larger scatter comparing with the as-received HSLA. There was obvious reduction in MPM of weld-repaired HSLA without BL, and in BL + MPM of weld-repaired HSLA with 10 mm BL. This is due to the formation of predominately block ferrite. The softening in HSLA welding joint was also reported in Ref. [12]. However, in the BL + MPM of weld-repaired HSLA with 4 mm BL, the fine-grained lath ferrite, bainite and pearlite greatly increased the VH number. This indicates that the incorporation of BL in weld-repaired HSLA removed the softening of welding joint, especially in the welding seam.
4.1. Effect of different welding procedures on the microstructure of welded interfaces Microstructure was examined using optical microscope with special emphasis on the regions around the welded interfaces. Fig. 6 displays the details of welded interfaces for the weldrepaired HSLA without BL, Fig. 7 displays the details of welded interfaces for the weld-repaired HSLA with 4 mm BL, and Fig. 8 displays the details of welded interfaces for the weld-repaired HSLA with 10 mm BL.
4.3. Comparison of the Paris fatigue curves of as-received and weldrepaired HSLA Because the Paris fatigue curves in each group were quite consistent, one specimen in each group was selected to compare. As shown in Fig. 10, da/dN from the weld-repaired HSLA with and without BL, fluctuated in the WM, MPM (BL + MPM) and HAZ, and had a larger scatter comparing with the as-received HSLA.
(a) MPM region of Weld-repaired HSLA without BL
(b) BL+MPM region of Weld-repaired
(c) BL+MPM region of Weld-repaired
HSLA with 4 mm BL
HSLA with 10 mm BL
Fig. 11. SEM micrographs in MPM or BL + MPM regions of weld-repaired HSLA with and without BLs.
242
C. Zhang et al. / Materials and Design 36 (2012) 233–242
As shown in Fig. 10, in the second stage of fatigue, weld-repaired HSLA without and with 10 mm BL show higher da/dN than that of PM, but weld-repaired HSLA with 4 mm BL shows lower da/ dN than that of PM. It should be mentioned that the lowest hardness in the MPM (weld-repaired HSLA without BL) and BL + MPM (weld-repaired HSLA with 10 mm BL) shown in Fig. 9 is also corresponding to the sudden jump of da/dN at the soft zones shown in Fig. 10. Similar findings had also been reported in Refs. [10,12]. Refs. [19,20] reported da/dN in the WM and HAZ of a weldrepaired HSLA without post weld heat treatment was much higher than that of the PM. But, with laser multiple-tempering or appropriate post weld heat treatment, the da/dN was improved to similar to that of the PM. However, the da/dN in weld-repaired HSLA with 4 mm BL was much lower than that of PM. This indicates that the moderate thick (4 mm) BL increased the VH number (Fig. 9) and fatigue resistance (Fig. 10) of welding joint. 4.4. Comparison of fatigue fractured surfaces of as-received and weldrepaired HSLA The VH curves and Paris fatigue curves of weld-repaired HSLA associated with various regions are shown in Figs. 3–5, but the most clearly difference occurred in MPM of weld-repaired HSLA without BL (Fig. 3, ‘‘3’’), and BL + MPM of weld-repaired HSLA with 4 mm (Fig. 4, ‘‘2’’) or 10 mm (Fig. 5, ‘‘3’’) BLs. The corresponding regions were performed using SEM, and are shown in Fig. 11. Fig. 11a and c display some spherical shaped non metallic particles in MPM/BL + MPM. This finding appears to be consistent with the lower VH number and fatigue resistance as shown by the VH and Paris fatigue measurements. The reason was the spherical shaped non metallic particles were harmful to fatigue behavior because of the stress concentration. Interestingly, the spherical shaped non metallic particles were eliminated by the incorporation of the BL with moderated thickness (4 mm), and the fatigue behavior was improved. 5. Conclusions Buffer layer (BL) between weld metal (WM) and parent metal (PM) was introduced to increase fatigue resistance of welding joint. The effect of microstructure of extensively weld-repaired HSLA steel with or without a BL on the mechanical properties of welding joint was studies in details. The following conclusions were drawn from the above study: 1) For the weld-repaired HSLA, the fatigue behavior and Vickers hardness (VH) of welding joint were closely interrelated with their microstructure. The mechanical properties of weld-repaired HSLA were increased by: increasing the amount of lath ferrite, pearlite and bainite; increasing fineness of grains structure. The block ferrite was harmful to fatigue resistance and improvement of welding softening. 2) For the weld-repaired HSLA with the specified material compositions and properties, the thin 4 mm BL worked well by virtually refining grain of all zones (WM, BL + MPM, heat affected zone-HAZ) of welding joint presented in the weldrepaired HSLA without BL. The forming lath ferrite, fine pearlite and batinte increased VH number and fatigue resistance of the welding joint. 3) The improvement in fatigue resistance was less significant for the weld-repaired HSLA with 10 mm BL as the formation of predominately block ferrite dropped the VH number and fatigue resistance.
4) The spherical shaped non-metallic particles greatly reduced the fatigue behavior of welding joint. 5) In the HAZs, VH was affected by their microstructures, but the fatigue crack growth rate was slightly influenced. In summary, a soft BL of a moderate thickness was beneficial as far as the fatigue behavior and hardness of weld-repaired HSLA was concerned according to the test results of the current study. However, excessive usage of a soft BL weakened a welded structure because of its low strength and the formation of block ferrite in the welding joint. Acknowledgements This work is supported by the Doctor Postgraduate Technical Project of Chang’an University (No. CHD2011ZY021). The authors (C.G. Zhang) wish to take this opportunity to express his gratitude. The authors would also like to thank Prof. Jin Xue from Xi’an Jiaotong University for his constructive comments during the course of this work. References [1] High strength low alloys steels. [20.08.11]. [2] Scholl M, Devanathan R, Clayton P. Abrasive and dry sliding wear resistance of Fe–Mo–Ni–Si and Fe–Mo–Ni–S–C weld hardfacing alloys. Wear 1990;135:355–68. [3] Bissalloy steels. Bisplate technical manual. P48. [22.09.06]. [4] Borozdin AV, Pavlov AA, Kroshkin VA. Experience with the use of high strength low alloy steels. Translated Khim Neft Mashinostr 1991:31–2. [5] Avazkonandeh-Gharavol MH, Haddad-Sabzevar M, Haerian A. Effect of copper content on the microstructure and mechanical properties of multipass MMA, low alloy steel weld metal deposits. Mater Des 2009;30:1902–12. [6] Beidokhti B, Koukabi AH, Dolati A. Influences of titanium and manganese on high strength low alloy SAW weld metal properties. Mater Charact 2009;60:225–33. [7] Lee S, Kim BC, Lee DY. Fracture mechanism in coarse grained HAZ of HSLA steel welds. Scr Metall 1989;23:995–1000. [8] Tsay LW, Chern TS, Gau CY, Yang JR. Microstructures and fatigue crack growth of EH36 TMCP steel weldments. Int J Fatigue 1999;21:857–64. [9] Krishnader M, Ghali E, Larouche M, Sridhar R, Lakshmanan VI. Cleavage failure of transformer storage tank under dynamic rates of loading: Influence of base plate and weldment microstructure and toughness. Eng Fail Anal 2006;13:1220–32. [10] Onoro J, Ranninger C. Fatigue behavior of laser welds of high-strength lowalloy steels. J Mater Process Technol 1997;68:68–70. [11] Omweg GM, Frankel GS, Bruce WA, Ramirez JE, Koch G. Performance of welded high-strength low-alloy steels in sour environments. Corrosion 2003;59:640–53. [12] Mohandas T, Reddy GM, Kumar BS. Heat-affected zone softening in highstrength low alloy steels. J Mater Process Technol 1999;88:284–94. [13] Mohandas T, Madhusudan-Reddy G, Satish-Kumar B. Heat-affected zone softening in high-strength low-alloy steels. J Mater Process Technol 1999;88:284–94. [14] Shi YW, Han ZX. Effect of weld thermal cycle on microstructure and fracture toughness of simulated heat-affected zone for a 800 MPa grade high strength low alloy steel. J Mater Process Technol 2008;207:30–9. [15] Ghosh M, Kumar K, Mishra RS. Friction stir lap welded advanced high strength steels: microstructure and mechanical properties. Mater Sci Eng, A 2011;528:8111–9. [16] Wan XL, Wei R, Wu KM. Effect of acicular ferrite formation on grain refinement in coarse-grained region of heat-affected zone. Mater Charact 2010;61:726–31. [17] Cwiek J. Hydrogen delayed cracking of high-strength weldable steels. Adv Mater Sci 2005;5:5–13. [18] Standard test method for measurement of fatigue crack growth rates. E647 (23). Annual Book of ASTM Standards. United States; 2001. [19] Tsay LW, Li YM, Chen C, Cheng SW. Mechanical properties and fatigue crack growth rate of laser-welded 4130 steel. Int J Fatigue 1992;14:239–47. [20] Ohta A, Sasaki E, Niher M, Kosuge M, Kanao M, Inagaki M. Fatigue crack propagation rates and threshold stress intensity factors for welded joints of HT80 steel at several stress ratios. Int J Fatigue 1982:233–7.
Contents lists available at SciVerse ScienceDirect
Materials and Design journal homepage: www.elsevier.com/locate/matdes
Effect of microstructure on mechanical properties in weld-repaired high strength low alloy steel Chunguo Zhang a,b,1, Xuding Song a, Pengmin Lu a,⇑,1, Xiaozhi Hu b a b
School of Construction Machinery, Chang’an University, Xi’an 710064, PR China School of Mechanical and Chemical Engineering, The University of Western Australia, Perth 6009, Australia
a r t i c l e
i n f o
Article history: Received 20 September 2011 Accepted 8 November 2011 Available online 15 November 2011 Keywords: D. Welding F. Microstructure G. Fractography
a b s t r a c t To understand the effect of microstructure on mechanical properties of weld-repaired high strength low alloy (HSLA), as-received and weld-repaired HSLA with and without buffer layers (BLs) were prepared. Microstructure analysis was carried out using optical microscope and SEM, and mechanical properties were measured by Vickers hardness test and fatigue test. The fatigue resistance of weld-repaired HSLA without BL was deteriorated with comparing to parent metal (PM). Meanwhile, Vickers hardness (VH) showed an obviously reduction in the melted parent metal (MPM), which was due to formation of predominately block ferrite. For the weld-repaired HSLA with BL, the VH and fatigue resistance increased with the incorporation of 4 mm BL, which was mainly due to formation of lath ferrite and fine-grained pearlite and bainite. When BL thickness increased to 10 mm, the VH and fatigue resistance decreased, which was because the thick BL diluted the MPM. VH number from low temperature (below melting point) heat affected zone (HAZ) fluctuated, but had a little scatter. However, the fatigue crack growth rate from HAZ was not obviously affected by the welding as comparison with the PM. Ó 2011 Elsevier Ltd. All rights reserved.
1. Introduction High strength low alloy (HSLA) steels with excellent strength characteristics have been widely used in various applications including cars, trucks and cable-stayed bridges. The strength of HSLA steels comes from their microstructure. Strength is increased by: increasing the amount of pearlite and bainite, increasing the fineness of the grains structure, increasing the amount of hard precipitate [1]. Wear damaged sections of supporting structures manufactured from HSLA are often weld-repaired or filled by similar metals through welding [2], or weld deposited by hard-facing alloys to provide a wear resistant surface for structural base materials. However, HSLA steels are hard to weld repair and special care is required because of their high strength [3,4]. Furthermore, the welding joints of HSLAs are often become the weak link in mechanical testing due to softening in the welding joints. To understand the detrimental influence of the welding on mechanical properties of welded HSLAs, a number of studies have been performed to
⇑ Corresponding author. Address: School of Construction Machinery, Chang’an University, Nan Er Huan Zhong Duan, Xi’an 710064, PR China. E-mail addresses: [email protected] (C. Zhang), [email protected] (X. Song), [email protected] (P. Lu), [email protected] (X. Hu). 1 Tel./fax: +86 29 82334322. 0261-3069/$ - see front matter Ó 2011 Elsevier Ltd. All rights reserved. doi:10.1016/j.matdes.2011.11.016
assess the effects of chemical compositions, metallographic structure, welding procedure and heat treatment (pre- and post-). With increasing of the copper content in weld metal, microstructure became finer in all zones of the weld metal, hardness and tensile strength increased, but Charpy impact toughness decreased [5]. The increase of manganese content was useful to refine and homogenize the weld microstructure, and the increase of titanium content in the range of 0.02–0.08% increased the acicular ferrite for the API 5L-X70 (HSLA) with moderate manganese content [6]. The amount of M-A constituent in coarse-grained HAZ of HSLA influenced the toughness value, and voids and micro-crack initiated at the M-A constituents [7]. The high-toughness low-carbon bainite in the HAZ of EH36 TMCP decreased fatigue crack growth rate, and post weld heat treatment was unnecessary [8]. Due to the coarse upper bainite, the weld metal was the weakest link in terms of toughness and fatigue resistance [9]. The fusion zone of laser welded HSLA, near to the interface of fusion zone and HAZ, was the weakest link of fatigue behavior [10]. Under the condition of gas metal arc welding, the high hardness welds (over 248HV) of HSLAs (API 5L X70 and X80) increased the resistance to sulfide stress cracking at low H2S concentrations [11]. Welded HSLA is often prone to softening in HAZ, which causes a weak link in mechanical properties [12]. However, HAZ of HSLA showed lesser softening in low heat input welding than in high heat input welding, and post weld heat treatment about AC3 and
234
C. Zhang et al. / Materials and Design 36 (2012) 233–242
external cooling were beneficial to reduce the tendency for softening in HAZ [13]. Reducing cooling time decreased the austenite grain size and volume fraction of M/A constituent, thus increased the toughness of HAZ of HSLA [14]. The cooling rate changed the microstructure in weld metal was also reported by Ghosh, et al. [15]. In the coarse-grained region of HAZ, the microstructure consisted of predominantly bainite packets and a small proportion of acicular ferrite was more stable than bainite during tempering [16]. Submerged arc welded joint of HSLA (14HNMBCu grade S690Q) had higher resistance to hydrogen degradation in sea water environment than shielded metal arc welded joint [17]. With different emphases to those aforementioned studies on welded HSLA steels, buffer layer between weld metal and parent metal was introduced to increase the fatigue resistance of welding joint in this paper. The primary objective of the present study was to investigate the metallographic structure of extensively weldrepaired HSLA steel with or without a buffer layer, and its influence on the mechanical properties of welding joint, which has not been systematically studied as shown in previous literatures.
imens’ surfaces after polishing and etching were obtained using professional digital camera. The micro-structure of fatigue surfaces of tested specimens associated with various regions in as-received and welded specimens, especially in MPM and BL + MPM, were observed in detail by a scanning electron microscopy (SEM).
Table 2 Mechanical properties of PM, WM and BL. Series
Tensile properties Yield strength (Mpa)
Tensile strength (Mpa)
Elongation (%)
Impact energy (J)
PM
690 min
790 min
18 min
WM
690 min
760 min
17 min
BL
410 min
490 min
22 min
40 min ( 20 °C) 50 min av. ( 51 °C) 27 min av. ( 29 °C)
2. Experimental procedure 2.1. Material and specimen The substrate material used for the present investigation was a HSLA steel (Bisplate80-parent metal PM) with a low carbon content. The weld metal (WM) and buffer layer (BL) were SmoothCor™ 115 and SmoothCor™ 70C6, respectively, thus effectively forming a welding joint together with PM with different strength characteristics. Flux cored arc welding was employed to perform the extensive welding repair, while 100% CO2 was used as the shielding gas. The core wire had a diameter of 1.2 mm, and electrode stick-out was 20 mm. A constant current power source (230A, 27 V) with DC + polarity was used in all cases (WM, 4 and 10 mm BLs). The chemical compositions and mechanical properties of PM, WM and BL are listed in Tables 1 and 2. Extended-Compact Tension (E-CT) specimens with machined through-the-thickness notches were machined following the specifications of ASTM E647 (23) [18]. The specimens were divided into four groups: (1) as-received HSLA, (2) weld-repaired HSLA without BL, (3) weld-repaired HSLA with 4 mm BL, and (4) weld-repaired HSLA with 10 mm BL. Dimensions of the E-CT specimens are shown in Fig. 1, and all dimensions are in mm.
(b) Weld-repaired HSLA without BL
(a) As-received HSLA (mm)
2.2. Microstructure study The four groups of E-CT specimens, (1) as-received HSLA specimens, (2) weld-repaired specimens without BL between the WM and PM, (3) weld-repaired specimens with 4 mm BL, and (4) weld-repaired specimens with 10 mm BL, were mechanically polished and etched using 4% nital solution for a few seconds to reveal their microstructures. The microstructures of the four groups of E-CT specimens were examined in detail by inverted optical microscope (JVC CMM-22E). The macrographs of the weld-repaired spec-
(d) Weld-repaired HSLA with 10 mm BL
(c) Weld-repaired HSLA with 4 mm BL
Fig. 1. Dimensions of E-CT specimens.
Table 1 Chemical compositions of PM, WM and BL. Series
PM WM BL
Element (wt.%) C
Si
Mn
Cr
Mo
Ni
P
S
B
CE(IIW)
0.18 0.06 0.03
0.20 0.30 0.59
1.40 1.40 1.66
0.20 0.22 –
0.20 0.44 –
– 2.29 –
0.01 – –
0.003 – –
0.0010 – –
0.50 0.58 0.31
CE(IIW) = C + Mn/6 + (Cu + Ni)/15 + (Cr + Mo + V)/5.
C. Zhang et al. / Materials and Design 36 (2012) 233–242
2.3. Mechanical properties The Vickers hardness variation form the WM, BL, melted parent metal (MPM) during welding, HAZ to PM in weld-repaired HSLA was measured with an indenting load of 20 kg by using a hardness tester (Mitutoyo AVK-C2). The four groups of E-CT specimens were tested under the same fatigue loading condition using Instron 8501. Constant amplitude loading with haversine waveform at a frequency of 5 Hz was used, and the R ratio was set at 0 throughout the test. The crack length was measured from the loading line as indicated in Fig. 1b. The Vickers hardness profiles and Paris fatigue curves for each group were measured twice with two identical specimens, and the results were quite consistent. 3. Results Microstructure of as-received and weld-repaired HSLA was examined using optical microscope with special emphasis on the regions in WM, MPM, HAZ, and around the welded interfaces and fusion line (FL). To correlate the microscopic observations to the relevant locations and corresponding VH and da/dN measurements, sketches for the boundary regions, VH distribution curves and da/dN curves were numbered accordingly for the weldrepaired HSLA with and without BLs. 3.1. As-received HSLA Following the above descriptions, Fig. 2 displays the details for PM. The VH distribution curves are shown in Fig. 2a, the results are quite consistent ranging from 224VH to 264VH, which can be as a good reference to study hardness variation in the weld-repaired HSLA with and without BLs. Paris fatigue curves from the as-
235
received HSLA are shown in Fig. 2b. Fig. 2c is an optical micrograph of the PM, showing a microstructure consisted of ferrite, bainite and pearlite. The VH measurements and fatigue propagation tests from two separate specimens were very consistent, which therefore provided a good reference to study effects of welding, incorporation of BL and various BL thicknesses. 3.2. Weld-repaired HSLA without BL Fig. 3 displays the macrograph, VH distribution curves, Paris fatigue curves, relevant locations and microstructure details for the weld-repaired HSLA without BL. The macrograph of surface after etching is shown in Fig. 3a, which displays an obviously white band of approximately 5 mm in width. The band was the melted parent metal (MPM) during welding. This was due to the carbon loss because of the over high temperature during the welding. As shown in Fig. 3f, the micrograph of MPM (white band) consisted of predominately block ferrite and some pearlite. The VH number in the corresponding position decreased. This is because carbon and alloy elements in MPM changed during welding. The microstructure of predominately block ferrite produced obvious reduction in VH shown in Fig. 3b, and greatly increase of fatigue crack propagation rate (da/dN) shown in Fig. 3c. Microstructures of distinct regions (WM, MPM and HAZ) in the weld-repaired HSLA without BL are shown in Fig. 3d–i. The microstructures of the WM in a multi-pass weld were dependent on the homogeneity of composition and thermal history of welding. The as-welded microstructures were reheated by the subsequent passes, thus different structures were obtained according the related thermal cycles and chemical compositions. The microstructure of the WM far from the MPM was consisted of predominately pearlite, and some ferrite (Fig. 3d designated by ‘‘1’’). As shown in
(a) VH distribution curves
(c) Microstructure Fig. 2. As-received HSLA.
(b) Paris fatigue curves
236
C. Zhang et al. / Materials and Design 36 (2012) 233–242
Fig. 3b and c (designated by ‘‘1’’), the fatigue resistance and VH number were increased by increasing the amount of pearlite. In the WM near the MPM, the amount of pearlite decreased with increasing the amount of ferrite (Fig. 3e designated by ‘‘2’’). The
microstructure reduced the VH number and fatigue strength shown in Fig. 3b and c (designated by ‘‘2’’), respectively. The microstructures of the HAZ were dependent on the temperature during welding. As shown in Fig. 3b, VH number fluctuated in
(a) Macrograph
(b) VH distribution curves
(c) Paris fatigue curves
(d) WM far from the MPM
(e) WM near to the MPM
(f) MPM
(g) Overheated zone in the HAZ
(h) Normalized zone in the HAZ
(i) Partial phase transformation zone in the HAZ
Fig. 3. Weld-repaired HSLA without BL.
C. Zhang et al. / Materials and Design 36 (2012) 233–242
the HAZ because of the different structure, and was reduced in overheated and partial phase transformation zones. The micro-
structure of overheated zone was predominately block ferrite and some bainite shown in Fig. 3g. The block ferrite was harmful to
(a) Macrograph
(b) VH distribution curves
237
(c) Paris fatigue curves
(d) WM
(e) BL+MPM
(f) Overheated zone in HAZ
(g) Normalized zone in HAZ
(h) Partial phase transformation zone in HAZ Fig. 4. Weld-repaired HSLA with 4 mm BL.
238
C. Zhang et al. / Materials and Design 36 (2012) 233–242
fatigue resistance. But the da/dN was similar to the PM shown in Fig. 3c, this is because the stress gradient reduced the da/dN as the crack propagated from soft MPM to strong HAZ. In the normalized zone, a significant grain refinement (ferrite + pearlite) was observed (Fig. 3h), which increased the VH number. The
microstructure of partial phase transformation zone consisted of coarse-grained ferrite together with some pearlite (Fig. 3i), which reduced the VH number. As shown in Fig. 3c, the da/dN from HAZ was similar to the PM.
(a) Macrograph
(b) VH distribution curves
(d) WM far from the BL+MPM
(c) Paris fatigue curves
(e) WM near the BL+MPM
(f) BL + MPM
(g) Overheated zone in the HAZ
(h) Normalized zone in the HAZ
(i) Partial phase transformation zone in the HAZ
Fig. 5. Weld-repaired HSLA with 10 mm BL.
C. Zhang et al. / Materials and Design 36 (2012) 233–242
3.3. Weld-repaired HSLA with 4 mm BL Fig. 4 displays the macrograph, VH distribution curves, Paris fatigue curves, relevant locations and microstructure details for the weld-repaired HSLA with 4 mm BL.
(a) The interface of WM and MPM
(c) The interface of MPM and HAZ
The macrograph of surface after etching is shown in Fig. 4a, which displays an obviously dark band of approximately 4.5 mm in width. The band was the mixture of BL and MPM during welding, which was called BL + MPM. Because both the BL and MPM were thin, the mixture was homogeneous. The microstructure of
(b) Enlarged view of rectangle in (a)
(d) Columnar crystal (rectangle in (c))
Fig. 6. Weld-repaired HSLA without BL.
(a) The interface of WM and BL+MPM
(c) The interface of BL+MPM and HAZ
239
(d) Columnar crystal (rectangle in (c))
Fig. 7. Weld-repaired HSLA with 4 mm BL.
240
C. Zhang et al. / Materials and Design 36 (2012) 233–242
(a) The interface of WM and BL+MPM
(c) The interface of BL+MPM and HAZ
(b) Enlarged view of rectangle in (a)
(d) Columnar crystal (rectangle in (c))
Fig. 8. Weld-repaired HSLA with 10 mm BL.
Fig. 9. Comparison of the VH distribution curves of weld-repaired HSLA with and without BL.
the BL + MPM was ferrite (predominately lath ferrite together with some granular and small block), fine pearlite and some bainite, shown in Fig. 4e. As shown in Fig. 4b and c, the predominately lath ferrite together with pearlite and bainite increased both VH number and fatigue resistance. In the WM and BL + MPM, the mixed microstructures composed of lath ferrite, granular ferrite and some block ferrite together with pearlite and bainite (Fig. 4d and e) were existed. Other optical metallographic results (not shown here) indicated there was no significant micro-structural change between the WM and BL + MPM. The microstructure is beneficial to increase of VH and fatigue behavior, which was proved by showing high VH number (Fig. 4b) and low da/dN (Fig. 4c). Microstructures of distinct regions in the HAZ are shown in Fig. 4f–h. The microstructure of the overheated and normalized zones was consisted of fine-grained ferrite and pearlite. The microstructure of the partial phase transformation was consisted of block ferrite and pearlite. As shown in Fig. 4b, VH number in the HAZ fluctuated, and there was reduction in overheated and partial phase transformation zones. The da/dN in the normalized and partial phase transformation zones was similar to the PM, but slightly higher in overheated zone than that of PM. This was because the stress gradient accelerated the fatigue crack growth as the crack propagated from strong BL + MPM to relative soft HAZ. 3.4. Weld-repaired HSLA with 10 mm BL
Fig. 10. Comparison of the Paris fatigue curves of weld-repaired HSLA with and without BL.
Fig. 5 displays the macrograph, VH distribution curves, Paris fatigue curves, relevant locations and microstructure details for the weld-repaired HSLA with 10 mm BL. The macrograph of surface after etching is shown in Fig. 5a, which displays an obviously white band of approximately 8 mm in width. The band was the mixture of BL and MPM during welding, which was called BL + MPM. The microstructure of the BL + MPM was predominately block ferrite together with a little pearlite, shown in Fig. 4f. As shown in Fig. 5b and c, the predominately block ferrite reduced VH number and fatigue resistance.
C. Zhang et al. / Materials and Design 36 (2012) 233–242
The microstructure of the WM far from the BL + MPM was consisted of predominately pearlite, and some ferrite (Fig. 5d designated by ‘‘1’’). As shown in Fig. 3b and c (designated by ‘‘1’’), the fatigue resistance and VH number were increased by increasing the amount of pearlite. In the WM near the MPM, the amount of pearlite decreased with increasing the amount of ferrite (Fig. 5e designated by ‘‘2’’). The microstructure reduced the VH number and fatigue strength shown in Fig. 5b and c (designated by ‘‘2’’), respectively. Microstructures of distinct regions in the HAZ are shown in Fig. 5g–i. The microstructure of the overheated zone was predominately widmanstatten structure. This was due to repeatedly heating during welding. The widmanstatten structure reduced the strength, and increased the brittleness. This was proved by showing high VH number (Fig. 5b) and high da/dN (Fig. 5c). As shown in Fig. 5h, the grain size of fine-grained normalized zone (pearlite + ferrite) was significantly smaller than that of PM. In the partial phase transformation zone, the microstructure was consisted of ferrite, and pearlite. The different metallographic structures of distinct regions in the HAZ changed the VH number, but did not influence the fatigue resistance obviously.
4. Discussions
241
The grain size around welded interfaces in weld-repaired HSLA with 4 mm BL (Fig. 7) was significantly smaller by showing lath ferrite structure than that in weld-repaired HSLA without BL (Fig. 6) because of the incorporation of 4 mm BL. In the case of 10 mm BL, grain refinement was not obviously (Fig. 8). The lath ferrite structure was beneficial to increase of fatigue resistance and VH number (Fig. 4b and c).
4.2. Analysis on Vickers hardness Because the VH distribution curves in each group were quite consistent, one specimen in each group was selected to compare. As shown in Fig. 9, VH number from the weld-repaired HSLA with and without BL, fluctuated in the WM, MPM (BL + MPM) and HAZ, and had a larger scatter comparing with the as-received HSLA. There was obvious reduction in MPM of weld-repaired HSLA without BL, and in BL + MPM of weld-repaired HSLA with 10 mm BL. This is due to the formation of predominately block ferrite. The softening in HSLA welding joint was also reported in Ref. [12]. However, in the BL + MPM of weld-repaired HSLA with 4 mm BL, the fine-grained lath ferrite, bainite and pearlite greatly increased the VH number. This indicates that the incorporation of BL in weld-repaired HSLA removed the softening of welding joint, especially in the welding seam.
4.1. Effect of different welding procedures on the microstructure of welded interfaces Microstructure was examined using optical microscope with special emphasis on the regions around the welded interfaces. Fig. 6 displays the details of welded interfaces for the weldrepaired HSLA without BL, Fig. 7 displays the details of welded interfaces for the weld-repaired HSLA with 4 mm BL, and Fig. 8 displays the details of welded interfaces for the weld-repaired HSLA with 10 mm BL.
4.3. Comparison of the Paris fatigue curves of as-received and weldrepaired HSLA Because the Paris fatigue curves in each group were quite consistent, one specimen in each group was selected to compare. As shown in Fig. 10, da/dN from the weld-repaired HSLA with and without BL, fluctuated in the WM, MPM (BL + MPM) and HAZ, and had a larger scatter comparing with the as-received HSLA.
(a) MPM region of Weld-repaired HSLA without BL
(b) BL+MPM region of Weld-repaired
(c) BL+MPM region of Weld-repaired
HSLA with 4 mm BL
HSLA with 10 mm BL
Fig. 11. SEM micrographs in MPM or BL + MPM regions of weld-repaired HSLA with and without BLs.
242
C. Zhang et al. / Materials and Design 36 (2012) 233–242
As shown in Fig. 10, in the second stage of fatigue, weld-repaired HSLA without and with 10 mm BL show higher da/dN than that of PM, but weld-repaired HSLA with 4 mm BL shows lower da/ dN than that of PM. It should be mentioned that the lowest hardness in the MPM (weld-repaired HSLA without BL) and BL + MPM (weld-repaired HSLA with 10 mm BL) shown in Fig. 9 is also corresponding to the sudden jump of da/dN at the soft zones shown in Fig. 10. Similar findings had also been reported in Refs. [10,12]. Refs. [19,20] reported da/dN in the WM and HAZ of a weldrepaired HSLA without post weld heat treatment was much higher than that of the PM. But, with laser multiple-tempering or appropriate post weld heat treatment, the da/dN was improved to similar to that of the PM. However, the da/dN in weld-repaired HSLA with 4 mm BL was much lower than that of PM. This indicates that the moderate thick (4 mm) BL increased the VH number (Fig. 9) and fatigue resistance (Fig. 10) of welding joint. 4.4. Comparison of fatigue fractured surfaces of as-received and weldrepaired HSLA The VH curves and Paris fatigue curves of weld-repaired HSLA associated with various regions are shown in Figs. 3–5, but the most clearly difference occurred in MPM of weld-repaired HSLA without BL (Fig. 3, ‘‘3’’), and BL + MPM of weld-repaired HSLA with 4 mm (Fig. 4, ‘‘2’’) or 10 mm (Fig. 5, ‘‘3’’) BLs. The corresponding regions were performed using SEM, and are shown in Fig. 11. Fig. 11a and c display some spherical shaped non metallic particles in MPM/BL + MPM. This finding appears to be consistent with the lower VH number and fatigue resistance as shown by the VH and Paris fatigue measurements. The reason was the spherical shaped non metallic particles were harmful to fatigue behavior because of the stress concentration. Interestingly, the spherical shaped non metallic particles were eliminated by the incorporation of the BL with moderated thickness (4 mm), and the fatigue behavior was improved. 5. Conclusions Buffer layer (BL) between weld metal (WM) and parent metal (PM) was introduced to increase fatigue resistance of welding joint. The effect of microstructure of extensively weld-repaired HSLA steel with or without a BL on the mechanical properties of welding joint was studies in details. The following conclusions were drawn from the above study: 1) For the weld-repaired HSLA, the fatigue behavior and Vickers hardness (VH) of welding joint were closely interrelated with their microstructure. The mechanical properties of weld-repaired HSLA were increased by: increasing the amount of lath ferrite, pearlite and bainite; increasing fineness of grains structure. The block ferrite was harmful to fatigue resistance and improvement of welding softening. 2) For the weld-repaired HSLA with the specified material compositions and properties, the thin 4 mm BL worked well by virtually refining grain of all zones (WM, BL + MPM, heat affected zone-HAZ) of welding joint presented in the weldrepaired HSLA without BL. The forming lath ferrite, fine pearlite and batinte increased VH number and fatigue resistance of the welding joint. 3) The improvement in fatigue resistance was less significant for the weld-repaired HSLA with 10 mm BL as the formation of predominately block ferrite dropped the VH number and fatigue resistance.
4) The spherical shaped non-metallic particles greatly reduced the fatigue behavior of welding joint. 5) In the HAZs, VH was affected by their microstructures, but the fatigue crack growth rate was slightly influenced. In summary, a soft BL of a moderate thickness was beneficial as far as the fatigue behavior and hardness of weld-repaired HSLA was concerned according to the test results of the current study. However, excessive usage of a soft BL weakened a welded structure because of its low strength and the formation of block ferrite in the welding joint. Acknowledgements This work is supported by the Doctor Postgraduate Technical Project of Chang’an University (No. CHD2011ZY021). The authors (C.G. Zhang) wish to take this opportunity to express his gratitude. The authors would also like to thank Prof. Jin Xue from Xi’an Jiaotong University for his constructive comments during the course of this work. References [1] High strength low alloys steels.