- Email: [email protected]
The relative effects of residual stresses and weld toe geometry on fatigue life of weldments
International Journal of Fatigue 77 (2015) 160–165
Contents lists available at ScienceDirect
International Journal of Fatigue journal homepage: www.elsevier.com/locate/ijfatigue
The relative effects of residual stresses and weld toe geometry on fatigue life of weldments Ebrahim Harati a,⇑, Leif Karlsson a, Lars-Erik Svensson a, Kamellia Dalaei b a b
Department of Engineering Science, University West, SE-461 86 Trollhättan, Sweden ESAB AB, Lindholmsallen 9, 40227 Gothenburg, Sweden
a r t i c l e
i n f o
Article history: Received 28 January 2015 Received in revised form 13 March 2015 Accepted 25 March 2015 Available online 2 April 2015 Keywords: Weld toe geometry Residual stress Fatigue strength Low Transformation Temperature filler material
a b s t r a c t The weld toe is one of the most probable fatigue crack initiation sites in welded components. In this paper, the relative influences of residual stresses and weld toe geometry on the fatigue life of cruciform welds was studied. Fatigue strength of cruciform welds produced using Low Transformation Temperature (LTT) filler material has been compared to that of welds produced with a conventional filler material. LTT welds had higher fatigue strength than conventional welds. A moderate decrease in residual stress of about 15% at the 300 MPa stress level had the same effect on fatigue strength as increasing the weld toe radius by approximately 85% from 1.4 mm to 2.6 mm. It was concluded that residual stress had a relatively larger influence than the weld toe geometry on fatigue strength. Ó 2015 Elsevier Ltd. All rights reserved.
1. Introduction Welded connections are the prime location for fatigue failures. Fatigue cracks usually initiate at the weld toe due to the local stress concentration caused by geometrical changes, residual stresses and also crack-like flaws that can sometimes be found in the weld toe [1–4]. Different improvement methods have been studied and implemented to increase the fatigue life of welded structures. The two main approaches are (a) modification of the weld toe geometry (by, for example, grinding or TIG re-melting) and (b) modification of the residual stress induced by welding (for example by shot peening or post weld heat treatment). The former improves the fatigue strength through reduction of the local stress concentration factor by ensuring a smooth transition between the weld profile and base metal. The latter contributes to fatigue strength increase by reducing tensile residual stresses or even by inducing compressive residual stresses at the weld toe region. There may be limitations for implementation of the mentioned techniques due to the technical complexity and also time and cost issues [5–7]. Among improvement methods, using Low Transformation Temperature (LTT) filler materials is a recent alternative approach to increase the fatigue strength of welded parts in a single step i.e., ⇑ Corresponding author. Tel.: +46 520 223341. E-mail address: [email protected] (E. Harati). http://dx.doi.org/10.1016/j.ijfatigue.2015.03.023 0142-1123/Ó 2015 Elsevier Ltd. All rights reserved.
without applying any costly post-weld treatment. A large number of investigations have confirmed the effectiveness of using LTT filler materials in increasing the fatigue life of welds. This is due to transformation of the weld at low temperatures, typically around 200 °C, which lowers the tensile residual stresses or even produces compressive residual stresses within the weld region and particularly at the weld toe [8–14]. However, LTT consumables are not widely commercially available and there is little experience from practical applications. LTT filler materials are hence not used extensively in industry. The effect of the weld geometry on fatigue strength of welded parts has been studied extensively. It has been found that the weld toe radius and angle, especially the weld toe radius, is one of the primary geometrical features that control the fatigue life. By increasing the weld toe radius, the angle or both radius and angle the fatigue strength will increase due to a decrease in stress concentration at the weld toe [15–22]. Thus, both residual stresses and geometrical features influence fatigue properties but, there is a lack of information about which has the stronger influence. Therefore, the aim of this paper is to evaluate the relative effects of weld toe geometry and residual stresses on fatigue life. In this regard fatigue strength of cruciform welds produced either using LTT or conventional filler materials has been compared. The effects of geometry and residual stresses at the weld toe have been evaluated and related to fatigue strength.
161
E. Harati et al. / International Journal of Fatigue 77 (2015) 160–165 Table 2 Mechanical properties of all-weld metal.
2. Material and methods 2.1. Base and filler materials
Welding consumable
R p 0.2 (MPa)
Rm (MPa)
A5 (%)
Impact toughness at 40 °C (J)
The base material used in this study was Weldox 700, a high strength steel with a plate thickness of 8 mm. The yield and tensile strengths of the base material were 817 MPa and 852 MPa, respectively. Gas Metal Arc Welding (GMAW) was used for cruciform welding of plates using two different filler materials, a standard high strength (OK Tubrod 14.03) and an LTT metal-cored wire. The compositions of the base material and filler materials are given in Table 1 and mechanical properties of filler materials are summarized in Table 2.
OK Tubrod 14.03a LTT
760 736
840 1127
23 13
70 49
a
Typical values.
this paper. Measurements were done at the centre as shown by the dashed line in Fig. 3a. Three independent measurements were done on the two opposite sides of the web. The evaluation was done at 20 magnification in a stereomicroscope and the result is the average of six measurements.
2.2. Welding setup 2.5. Residual stress measurements Four series of welds (L1, L2, C1, C2) were produced, each consisting of 11 samples. Of these, 10 samples were used for fatigue testing and one for geometry and residual stress measurements. L1 and C1 were single-pass welds while L2 and C2 were doublepass welds. L refers to welds produced by the LTT filler material and C refers to the welds produced by the conventional filler material (OK Tubrod 14.03). Ar + 2% CO2 was used as shielding gas. The welding parameters are summarized in Table 3. Dimensions and the configuration of cruciform test samples together with the welding sequence are shown in Fig. 1. In order to avoid the influence of welding start/stop positions on fatigue properties, they were located on each side of the flanges. One start/stop position is shown in the photograph of the test weld in Fig. 2. 2.3. Fatigue testing The fatigue testing was performed by applying a constant amplitude sinusoidal tensile load in the longitudinal direction, with a frequency of 29–40 Hz at a stress ratio R = 0.1. Testing was stopped before fracture (run-outs) for samples surviving more than 2 million cycles [23]. For evaluation of the test data, the number of cycles (N) was treated as the independent parameter in log(N) = log(C) m ⁄ log(Dr). In this equation, C and m are fitting constants. The number of cycles were plotted against the stress ranges (Dr), and using the least square technique, the mean fatigue strength was estimated [24]. To calculate the characteristic fatigue strength at 2 million cycles, the FAT value, the mean curve is lowered by k standard deviations. Here, a k value of 2.7 was used, based on the number of weld samples [25]. 2.4. Weld toe radius and angle measurements There are several destructive and non-destructive methods for measuring the weld toe radius and angle [15,26–28]. In this paper, the Weld Impression Analysis (WIA) method was used. Detailed information about using this method for measuring the weld toe radius has been published elsewhere [29]. Fig. 3 schematically presents the two geometrical features, i.e. weld toe angle (H) and weld toe radius (r), that were measured for the weld samples in
Surface residual stresses were measured using the X-ray sin2 u method. An XSTRESS 3000 X-ray Stress Analyzer, with Cr-Ka radiation and varying u – angles between 45° and +45° was used to estimate the stresses transverse to the weld toe. The aperture diameter was 2 mm. The stress was measured at the last weld along the centreline of the flanges at 1–40 mm from the weld toe [23]. Measurements were performed on the as received welded plates, with no prior grinding or polishing. 3. Results 3.1. Fatigue testing In all samples fracture initiated at the weld toes at the short edges of the flanges and propagated through the base metal. The fatigue test results are compiled in S–N curves shown in Fig. 4. For comparison, the IIW FAT-71 class [25] for the used specimen geometry is also presented in the figure. The calculated characteristic fatigue strength (FAT), the mean fatigue strength at 2 million cycles and the slope of the curves are tabulated in Table 4. It is clear that fatigue strength of LTT welds are significantly higher than for welds produced using the conventional wire (Fig. 4 and Table 4). The fatigue strength of the double-pass LTT welds is the highest, followed by the single-pass LTT welds and finally the single and double-pass welds produced by the conventional wire. From the FAT values in Table 4 it can be seen that a fatigue strength improvement of 22% and 55% is achieved for L1 and L2 welds, respectively, when compared with the IIW FAT value of 71 MPa. Experimental FAT-values for welds produced by the conventional filler material were slightly below the IIW FAT 71. Comparing mean fatigue strength levels at 2 million cycles even more clearly shows the higher fatigue strength of the LTT welds. A direct comparison shows an improvement of 58% for the single-pass welds and 76% for the double-pass welds. As can be seen in Fig. 4, the fatigue strength is almost similar for all consumables at high stresses, but higher for LTT consumables at lower loads. The slope of the S–N lines are 3.74 and 4.41 (Table 4) for welds by LTT fillers. These are higher than the slope of 3 used
Table 1 Chemical composition of base material and filler materials (weight %).
Weldox 700 LTT OK Tubrod 14.03a a
Nominal composition.
C
Si
Mn
P
S
Cr
Ni
Mo
V
Ti
Cu
Al
Nb
0.13 0.02 0.07
0.3 0.7 0.6
1.18 1.3 1.7
0.011 0.005 –
0.003 0.007 –
0.27 12.8 –
0.04 6.2 2.3
0.13 0.1 0.6
0.007 – –
0.013 – –
0.01 – –
0.041 – –
0.022 – –
162
E. Harati et al. / International Journal of Fatigue 77 (2015) 160–165 Table 3 Welding parameters. Weld
Filler material
Number of beads
Current (A)
Voltage (V)
Welding speed (mm/s)
Heat input (kJ/mm)
L1 L2 C1 C2
LTT LTT OK Tubrod 14.03 OK Tubrod 14.03
1 2 1 2
285 280 240 250
23.9 23.9 23.5 20.5
6.5 10 5.2 8.8
1.05 0.67 1.09 0.58
Fig. 1. Design and dimensions of the cruciform test specimens (dimensions in mm).
Fig. 2. Photograph of a test specimen.
for the FAT 71 curve and also higher than those of the conventional welds at 2.42 and 2.79. This results in significantly higher fatigue strength for LTT fillers at lower stresses.
Fig. 4. Fatigue test results for cruciform specimens produced using LTT (L1 and L2) and conventional filler materials (C1 and C2). Testing was done with constant amplitude loading at R = 0.1.
3.2. Residual stresses Fig. 5 shows the longitudinal surface residual stresses as function of the distance from the weld toe. The measurements were made in front of weld No. 4 in Fig. 1. In general the measured tensile residual stresses are higher for welds produced with the conventional wire compared to those welded with the LTT wire. The difference is most pronounced very close to the weld toe. For example, the tensile residual stress were almost 200 MPa lower, 1 mm from the weld toe, for single-pass samples welded using LTT compared to single-pass welds using the conventional wire.
Table 4 Characteristic fatigue strength (FAT), mean fatigue strength at 2 million cycles and slope of the S–N curves (m). Weld
FAT (MPa)
Mean fatigue strength (MPa)
m
L1 L2 C1 C2
87 110 65 69
117 137 74 78
3.74 4.41 2.42 2.79
Fig. 3. (a) Weld sample (b) Schematic geometrical features at the weld toe.
163
E. Harati et al. / International Journal of Fatigue 77 (2015) 160–165
Fig. 5. Longitudinal surface residual stresses (stresses transverse to the weld toe) in front of the weld toe for samples welded with LTT and conventional filler materials. Note the lower residual stresses near the weld toe for LTT welds.
It can also be seen that tensile residual stresses are lower at 1 mm compared to 2 mm from the weld toe for LTT welds, whereas the opposite is true for the conventional welds. At distances closer than 3 mm to the weld toe, residual stress is slightly higher for a double-pass LTT weld compared to the single-pass LTT. The trend is opposite at distances further away. For welds made using the conventional wire higher tensile residual stresses were found up to 8 mm from the weld toe for the single-pass samples, compared to double-pass welds. Further away from the weld toe they show almost the same values. When comparing the two welds produced with conventional consumables and the two welds produced with LTT consumables it should be noted that the differences within the two pairs are not the same, close to the weld toe. The difference between the two LTT welds is about 30 MPa. The difference between the conventional welds is about 40–50 MPa. The experimental uncertainty, close to the weld toe, for all welds is in the order of 4– 10 MPa and it reaches to a maximum of about 20 MPa at distances further away. 3.3. Weld toe radius and angle Table 5 shows the results of the weld toe radius and angle measurements using the WIA method. The largest values for both the weld toe radius and angle is found for the single-pass weld produced by conventional filler material (C1). No significant difference can be seen between L1, L2, and C2. The maximum and minimum values for toe radii and angles in Table 5 also show that C1 has the highest values but the ranking between the other welds is uncertain. 4. Discussion The aim of this paper is to compare the relative influence of toe geometry and residual stress on the fatigue life of cruciform welded joints. The results of the testing presented in the results section certainly point to that residual stress has a relatively larger influence than toe geometry. This is based on the fact that a significantly higher mean fatigue strength value was obtained for welds L1 and L2 than for the welds made using conventional filler materials. In particular weld C1, having a significantly larger weld toe radius than the other welds, would be expected to have a higher
mean fatigue strength if the weld toe radius were the most influential parameter. An interesting comparison of the relative effects of stress and toe radius can be made for the two welds produced with the conventional filler material (C1 and C2). These welds show almost identical mean fatigue strength, but C1 has a toe radius of 2.6 mm and C2 a radius of 1.4 mm (Table 5). Assuming the same difference in residual stress at the weld toe as 1 mm away, as results by Ramjaun et al. [30] suggest is reasonable for conventional welding consumables, this makes a quantitative comparison possible. A decrease in residual stress by about 40 MPa (Fig. 5) would then balance the decrease in toe radius from 2.6 mm to 1.4 mm. This result can be compared to those by Björk et. al. [31] who discussed the relationship between the weld toe geometry and fatigue strength for non-load carrying X-joints. From Fig. 15 in their paper a clear trend can be seen that the fatigue strength increases with increasing weld toe radius. Although there is significant scatter in the data, typically a 1 mm increase in the weld toe radius from 1.5 mm to 2.5 mm resulted in about 20–30 MPa increase in FAT. Although speculative this suggests that FAT of weld C1 should have been approximately 24–36 MPa higher than for C2 if residual stresses were similar. In other words it seems that a decrease in residual stress of 40 MPa should correspond to an increase on FAT of about 24–36 MPa in this particular case. This can be compared with results of studies by Nguyen and Wahab [32,33] who developed a mathematical model to investigate the effect of weld geometry and residual stresses on fatigue life of butt welds. The geometry is different from the present study but results suggest that an increase of about 2 mm in toe radius has approximately the same effect on fatigue strength as a decrease in residual stress of about 60 MPa. In other words roughly the same relation as found for the welds in the present study. Another observation is that welds L1 and L2 have almost identical weld toe radius and L2 has slightly higher residual stress, at least close to the weld toe. Still, the mean fatigue strength of L2 is significantly higher than for L1. If the residual stress was the most important factor, then a slightly lower mean fatigue strength would have been expected for L2. This seemingly contradictory result can most likely be explained in terms of the location of residual stress measurement and the technique used. Very recent results obtained by Ramjaun et al. [30] comparing the near surface (0.15 mm below surface) weld toe residual stresses for butt welds produced with conventional and LTT consumables using neutron and X-ray diffraction show interesting results. Very similar stresses were measured by X-ray and neutron diffraction at some distance from the weld toe. However, it was also shown that the X-ray method is not capable of capturing a dramatic local drop in stresses that occur within about 1 mm around the weld toe for LTT welds. For example both X-ray and neutron diffraction measured a transverse stress (normal to the weld line) of about 200 MPa at 1– 5 mm from the weld toe. With neutron diffraction it was also seen that there was a dramatic local change in stress to about 600 MPa at the weld toe fusion boundary. This is the local stress that is more likely to be relevant for fatigue properties than the stress measured at a distance 1 mm from the weld toe. No such change was observed for welds produced with conventional consumables where stresses were fairly constant within a few millimetres of
Table 5 Results from weld toe radius and angle measurements, Average from six measurements together with the maximum and minimum values. Weld
Radius (mm)
Maximum radius (mm)
Minimum radius (mm)
Angle (°)
Maximum angle (°)
Minimum angle (°)
L1 L2 C1 C2
1.4 1.3 2.6 1.4
1.5 1.6 3.1 1.7
1.2 1.0 2.2 1.0
128 134 149 119
145 136 159 140
127 131 148 120
164
E. Harati et al. / International Journal of Fatigue 77 (2015) 160–165
the weld toe. Therefore, X-ray measurement cannot be expected to accurately measure the actual local weld toe stress for LTT welds but can give an indication of whether the residual stresses are higher or lower than for conventional welds. It can also be concluded that they are most likely significantly lower than measured by X-ray at 1 mm. The higher fatigue strength of L2 in comparison to L1 can also be understood in the light of results from the same study by Ramjaun et. al. [30]. They found that the volume of the LTT weld metal has an effect on the local weld toe stress level. For a three-pass LTT weld compressive stresses of about 600 MPa were measured whereas the compressive stresses at a single-pass LTT weld was about 400 MPa. In both cases stresses next to the weld was about 200 MPa with the average stresses somewhat less compressive for the single pass weld. There are obviously geometrical differences between the welds in this paper and the referenced study that may result in different stress levels but it is likely that the overall behaviour is similar. Based on this it is likely that the greater volume of the double-pass weld L2 has contributed to a larger compressive weld toe stress than for the single-pass weld L1 even though X-ray measurements are not capable of resolving this. Furthermore the second weld pass of L2 will have a composition with a lower Ms-temperature due to less dilution and thereby more effectively lowering the residual stress level. Considering the above discussion, the much better fatigue properties of weld L2 than for L1 can therefore most likely be explained in terms of a lower residual stress locally at the weld toe. In conclusion the present results suggest that a moderate change in residual stress of about 15% at the 300 MPa stress level has the same effect on fatigue strength as the increase of the weld toe radius by approximately 85%. It should be noted that this result was obtained for radii ranging from 1.4 mm to 2.6 mm and it cannot be concluded from the present study whether the relation is linear and applies to other weld toe radii or stress ranges or not. The large effect of using LTT consumables than having a larger weld toe radius also shows the importance of the residual stress level. Certainly there is as always significant scatter in the fatigue data, variation in the toe radius along the weld and uncertainties in the measurement of the local stress at the weld toe. Nevertheless the results indicate that a decrease of the residual stress has relatively seen a larger effect on fatigue life than a change in weld toe radius. This is in many way quite reassuring as it is well known, and inevitably so, that there will always be significant variations in the local weld toe radius along the weld as seen in Table 5. Further studies are needed to make a more quantitative comparison and to study whether the effect is the same or varies for different stress levels and toe radii.
For welds made with conventional consumables, a change in residual stress of about 15% at the 300 MPa stress level had the same effect on fatigue strength as the increase of the weld toe radius by approximately 85% from 1.4 mm to 2.6 mm. Residual stress has a relatively larger influence than the weld toe geometry on fatigue strength of cruciform welds.
Acknowledgements The financial support of the Swedish Energy Agency is gratefully acknowledged. We thank ESAB AB for providing the welding filler materials and welding of the samples. The authors would also like to thank Kjell Hurtig (University West, Sweden) for his help with geometrical measurements. References [1] [2] [3]
[4]
[5]
[6]
[7]
[8]
[9]
[10]
[11]
[12]
[13]
[14]
5. Conclusions The fatigue strength of cruciform welds produced with conventional and with LTT filler materials was determined. By comparing welds with different toe radii and residual stress levels the following can be concluded: The fatigue strength of LTT welds at 2 million cycles was 58% higher for single-pass welds and 76% higher for the double-pass welds when compared to corresponding welds made with conventional filler materials. Surface residual stresses very close to the weld toe are higher for welds produced with the conventional wire compared to those welded with the LTT wire. The higher fatigue strength of LTT welds than for welds made with conventional consumables is an effect of the significantly lower stress levels at the weld toe.
[15]
[16] [17]
[18]
[19]
[20]
Stephens RI, Fatemi A, Stephens RR, Fuchs HO. Metal fatigue in engineering. NY: John Willey & Sons; 2001. Pang HLJ. Analysis of weld toe profiles and weld toe cracks. Int J Fatigue 1993;15:31–6. http://dx.doi.org/10.1016/0142-1123(93)90074-Z. Alam MM, Barsoum Z, Jonsén P, Häggblad HA, Kaplan A. Geometrical aspects of the fatigue behaviour of laser hybrid fillet welds. In: Proc. Fatigue Des Conf Senlis Paris Fr; 2009. Atzori B, Lazzarin P, Meneghetti G, Ricotta M. Fatigue design of complex welded structures. Int J Fatigue 2009;31:59–69. http://dx.doi.org/10.1016/ j.ijfatigue.2008.02.013. Kirkhope KJ, Bell R, Caron L, Basu RI, Ma K-T. Weld detail fatigue life improvement techniques. Part 1: Review. Mar Struct 1999;12:447–74. http:// dx.doi.org/10.1016/S0951-8339(99)00013-. Tveiten B, Fjeldstad A, Harkegard G, Myhr O, Bjorneklett B. Fatigue life enhancement of aluminium joints through mechanical and thermal prestressing. Int J Fatigue 2006;28:1667–76. http://dx.doi.org/10.1016/ j.ijfatigue.2006.01.006. Yildirim HC, Marquis GB. Fatigue strength improvement factors for high strength steel welded joints treated by high frequency mechanical impact. Int J Fatigue 2012;44:168–76. Çam G, Özdemir O, Koçak M. Progress in low transformation temperature (ltt) filler wires: review. In: Proc 63rd Annu Assem Int Conf Int Inst Weld, 2010; p. 759–65. Altenkirch J, Gibmeier J, Kromm A, Kannengiesser T, Nitschke-Pagel T, Hofmann M. In situ study of structural integrity of low transformation temperature (LTT)-welds. Mater Sci Eng, A 2011;528:5566–75. http:// dx.doi.org/10.1016/j.msea.2011.03.091. Kromm MA, Kannengiesser T. Characterizing phase transformations of different LTT alloys and their effect on residual stresses and cold cracking. Weld World 2011;55:48–56. http://dx.doi.org/10.1007/BF03321286. Bhatti AA, Barsoum Z, van der Mee V, Kromm A, Kannengiesser T. Fatigue strength improvement of welded structures using new low transformation temperature filler materials. Proc Eng 2013;66:192–201. http://dx.doi.org/ 10.1016/j.proeng.2013.12.07. Takahashi M, Yasuda HY. Variant selection of martensites in steel welded joints with low transformation temperature weld metals. J Alloys Compd 2013;577:S601–4. http://dx.doi.org/10.1016/j.jallcom.2012.02.022. Ooi SW, Garnham JE, Ramjaun TI. Review: low transformation temperature weld filler for tensile residual stress reduction. Mater Des 2014;56:773–81. http://dx.doi.org/10.1016/j.matdes.2013.11.050. Ramjaun TI, Stone HJ, Karlsson L, Kelleher J, Ooi SW, Dalaei K, et al. Effects of dilution and baseplate strength on stress distributions in multipass welds deposited using low transformation temperature filler alloys. Sci Technol Weld Join 2014;19:461–7. http://dx.doi.org/10.1179/1362171814Y.0000000209. Williams HE, Ottsen H, Lawence FV, Munse WH. The effects of weld geometry on the fatigue behavior of welded connrctions. University of Illinois Engineering Experiment Station. College of Engineering. University of Illinois at Urbana-Champaign.; 1970. Niu X, Glinka G. The weld profile effect on stress intensity factors in weldments. Int J Fract 1987;35:3–20. http://dx.doi.org/10.1007/BF00034531. Martins Ferreira JA, Moura Branco CA. Influence of the radius of curvature at the weld toe in the fatigue strength of fillet welded joints. Int J Fatigue 1989;11:29–36. http://dx.doi.org/10.1016/0142-1123(89)90044-3. Teng T-L, Fung C-P, Chang P-H. Effect of weld geometry and residual stresses on fatigue in butt-welded joints. Int J Press Vessels Pip 2002;79:467–82. http://dx.doi.org/10.1016/S0308-0161(02)00060-1. Caccese V, Blomquist PA, Berube KA, Webber SR, Orozco NJ. Effect of weld geometric profile on fatigue life of cruciform welds made by laser/GMAW processes. Mar Struct 2006;19:1–22. http://dx.doi.org/10.1016/ j.marstruc.2006.07.002. Lee C-H, Chang K-H, Jang G-C, Lee C-Y. Effect of weld geometry on the fatigue life of non-load-carrying fillet welded cruciform joints. Eng Fail Anal 2009;16:849–55. http://dx.doi.org/10.1016/j.engfailanal.2008.07.004.
E. Harati et al. / International Journal of Fatigue 77 (2015) 160–165 [21] Radhi HE, Barrans S. Finite Element Analysis of Effect of Weld Toe Radius and Plate Thickness on Fatigue Life of Butt Welded Joint. In: Lucas G, Xu Z, editors. Future Technol. Comput. Eng. Proc. Comput. Eng. Annu. Res. Conf. 2010 CEARC’10, Huddersfield: University of Huddersfield; 2010, p. 60–4. [22] Barsoum Z, Jonsson B. Influence of weld quality on the fatigue strength in seam welds. Eng Fail Anal 2011;18:971–9. http://dx.doi.org/10.1016/j.engfailanal. 2010.12.001. [23] Svensson L-E, Karlsson L, Harati E. Increasing fatigue life using Low Transformation Temperature (LTT) welding consumables. In: Proceeding of 2nd Swedish conference on design and fabrication of welded structures, Borlänge, Sweden; 2013, p. 49–64. [24] Dowling NE. Mechanical behavior of materials : engineering methods for deformation, fracture, and fatigue. Prentice Hall; 1993. [25] Hobbacher A. Recommendations for fatigue design of welded joints and components. International Institute of Welding; XIII-2460-13/XV-1440-13, Paris, France; 2014. [26] Hou C. Fatigue analysis of welded joints with the aid of real three-dimensional weld toe geometry. Int J Fatigue 2007;29:772–85. http://dx.doi.org/10.1016/ j.ijfatigue.2006.06.007. [27] Peng T. Algorithms and models for 3-D shape measurement using digital fringe projections. Doctoral theis: University of Maryland, College Park, USA; 2007.
165
[28] Stenberg T, Lindgren E, Barsoum Z. Development of an algorithm for quality inspection of welded structures. Proc Inst Mech Eng Part B J Eng Manuf 2012;226:1033–41. http://dx.doi.org/10.1177/0954405412439138. [29] Harati E, Ottosson M, Karlsson L, Svensson L-E. Non-destructive measurement of weld toe radius using Weld Impression Analysis, Laser Scanning Profiling and Structured Light Projection methods. In: First Int Conf Weld Non Destr Test Tehran, Iran; 2014. [30] Ramjaun TI, Stone HJ, Karlsson L, Gharghouri MA, Dalaei K, Moat RJ, et al. Surface residual stresses in multipass welds produced using low transformation temperature filler alloys. Sci Technol Weld Join 2014;19:623–30. http://dx.doi.org/10.1179/1362171814Y.0000000234. [31] Björk T, Samuelsson J, Marquis G. The need for a weld quality system for fatigue loaded structures. Weld World 2008;52:34–46. http://dx.doi.org/ 10.1007/BF03266615. [32] Ninh NT, Wahab MA. The effect of residual stresses and weld geometry on the improvement of fatigue life. J Mater Process Technol 1995;48:581–8. http:// dx.doi.org/10.1016/0924-0136(94)01697-Y. [33] Nguyen TN, Wahab MA. The effect of weld geometry and residual stresses on the fatigue of welded joints under combined loading. J Mater Process Technol 1998;77:201–8. http://dx.doi.org/10.1016/S0924-0136(97)00418-4.
Contents lists available at ScienceDirect
International Journal of Fatigue journal homepage: www.elsevier.com/locate/ijfatigue
The relative effects of residual stresses and weld toe geometry on fatigue life of weldments Ebrahim Harati a,⇑, Leif Karlsson a, Lars-Erik Svensson a, Kamellia Dalaei b a b
Department of Engineering Science, University West, SE-461 86 Trollhättan, Sweden ESAB AB, Lindholmsallen 9, 40227 Gothenburg, Sweden
a r t i c l e
i n f o
Article history: Received 28 January 2015 Received in revised form 13 March 2015 Accepted 25 March 2015 Available online 2 April 2015 Keywords: Weld toe geometry Residual stress Fatigue strength Low Transformation Temperature filler material
a b s t r a c t The weld toe is one of the most probable fatigue crack initiation sites in welded components. In this paper, the relative influences of residual stresses and weld toe geometry on the fatigue life of cruciform welds was studied. Fatigue strength of cruciform welds produced using Low Transformation Temperature (LTT) filler material has been compared to that of welds produced with a conventional filler material. LTT welds had higher fatigue strength than conventional welds. A moderate decrease in residual stress of about 15% at the 300 MPa stress level had the same effect on fatigue strength as increasing the weld toe radius by approximately 85% from 1.4 mm to 2.6 mm. It was concluded that residual stress had a relatively larger influence than the weld toe geometry on fatigue strength. Ó 2015 Elsevier Ltd. All rights reserved.
1. Introduction Welded connections are the prime location for fatigue failures. Fatigue cracks usually initiate at the weld toe due to the local stress concentration caused by geometrical changes, residual stresses and also crack-like flaws that can sometimes be found in the weld toe [1–4]. Different improvement methods have been studied and implemented to increase the fatigue life of welded structures. The two main approaches are (a) modification of the weld toe geometry (by, for example, grinding or TIG re-melting) and (b) modification of the residual stress induced by welding (for example by shot peening or post weld heat treatment). The former improves the fatigue strength through reduction of the local stress concentration factor by ensuring a smooth transition between the weld profile and base metal. The latter contributes to fatigue strength increase by reducing tensile residual stresses or even by inducing compressive residual stresses at the weld toe region. There may be limitations for implementation of the mentioned techniques due to the technical complexity and also time and cost issues [5–7]. Among improvement methods, using Low Transformation Temperature (LTT) filler materials is a recent alternative approach to increase the fatigue strength of welded parts in a single step i.e., ⇑ Corresponding author. Tel.: +46 520 223341. E-mail address: [email protected] (E. Harati). http://dx.doi.org/10.1016/j.ijfatigue.2015.03.023 0142-1123/Ó 2015 Elsevier Ltd. All rights reserved.
without applying any costly post-weld treatment. A large number of investigations have confirmed the effectiveness of using LTT filler materials in increasing the fatigue life of welds. This is due to transformation of the weld at low temperatures, typically around 200 °C, which lowers the tensile residual stresses or even produces compressive residual stresses within the weld region and particularly at the weld toe [8–14]. However, LTT consumables are not widely commercially available and there is little experience from practical applications. LTT filler materials are hence not used extensively in industry. The effect of the weld geometry on fatigue strength of welded parts has been studied extensively. It has been found that the weld toe radius and angle, especially the weld toe radius, is one of the primary geometrical features that control the fatigue life. By increasing the weld toe radius, the angle or both radius and angle the fatigue strength will increase due to a decrease in stress concentration at the weld toe [15–22]. Thus, both residual stresses and geometrical features influence fatigue properties but, there is a lack of information about which has the stronger influence. Therefore, the aim of this paper is to evaluate the relative effects of weld toe geometry and residual stresses on fatigue life. In this regard fatigue strength of cruciform welds produced either using LTT or conventional filler materials has been compared. The effects of geometry and residual stresses at the weld toe have been evaluated and related to fatigue strength.
161
E. Harati et al. / International Journal of Fatigue 77 (2015) 160–165 Table 2 Mechanical properties of all-weld metal.
2. Material and methods 2.1. Base and filler materials
Welding consumable
R p 0.2 (MPa)
Rm (MPa)
A5 (%)
Impact toughness at 40 °C (J)
The base material used in this study was Weldox 700, a high strength steel with a plate thickness of 8 mm. The yield and tensile strengths of the base material were 817 MPa and 852 MPa, respectively. Gas Metal Arc Welding (GMAW) was used for cruciform welding of plates using two different filler materials, a standard high strength (OK Tubrod 14.03) and an LTT metal-cored wire. The compositions of the base material and filler materials are given in Table 1 and mechanical properties of filler materials are summarized in Table 2.
OK Tubrod 14.03a LTT
760 736
840 1127
23 13
70 49
a
Typical values.
this paper. Measurements were done at the centre as shown by the dashed line in Fig. 3a. Three independent measurements were done on the two opposite sides of the web. The evaluation was done at 20 magnification in a stereomicroscope and the result is the average of six measurements.
2.2. Welding setup 2.5. Residual stress measurements Four series of welds (L1, L2, C1, C2) were produced, each consisting of 11 samples. Of these, 10 samples were used for fatigue testing and one for geometry and residual stress measurements. L1 and C1 were single-pass welds while L2 and C2 were doublepass welds. L refers to welds produced by the LTT filler material and C refers to the welds produced by the conventional filler material (OK Tubrod 14.03). Ar + 2% CO2 was used as shielding gas. The welding parameters are summarized in Table 3. Dimensions and the configuration of cruciform test samples together with the welding sequence are shown in Fig. 1. In order to avoid the influence of welding start/stop positions on fatigue properties, they were located on each side of the flanges. One start/stop position is shown in the photograph of the test weld in Fig. 2. 2.3. Fatigue testing The fatigue testing was performed by applying a constant amplitude sinusoidal tensile load in the longitudinal direction, with a frequency of 29–40 Hz at a stress ratio R = 0.1. Testing was stopped before fracture (run-outs) for samples surviving more than 2 million cycles [23]. For evaluation of the test data, the number of cycles (N) was treated as the independent parameter in log(N) = log(C) m ⁄ log(Dr). In this equation, C and m are fitting constants. The number of cycles were plotted against the stress ranges (Dr), and using the least square technique, the mean fatigue strength was estimated [24]. To calculate the characteristic fatigue strength at 2 million cycles, the FAT value, the mean curve is lowered by k standard deviations. Here, a k value of 2.7 was used, based on the number of weld samples [25]. 2.4. Weld toe radius and angle measurements There are several destructive and non-destructive methods for measuring the weld toe radius and angle [15,26–28]. In this paper, the Weld Impression Analysis (WIA) method was used. Detailed information about using this method for measuring the weld toe radius has been published elsewhere [29]. Fig. 3 schematically presents the two geometrical features, i.e. weld toe angle (H) and weld toe radius (r), that were measured for the weld samples in
Surface residual stresses were measured using the X-ray sin2 u method. An XSTRESS 3000 X-ray Stress Analyzer, with Cr-Ka radiation and varying u – angles between 45° and +45° was used to estimate the stresses transverse to the weld toe. The aperture diameter was 2 mm. The stress was measured at the last weld along the centreline of the flanges at 1–40 mm from the weld toe [23]. Measurements were performed on the as received welded plates, with no prior grinding or polishing. 3. Results 3.1. Fatigue testing In all samples fracture initiated at the weld toes at the short edges of the flanges and propagated through the base metal. The fatigue test results are compiled in S–N curves shown in Fig. 4. For comparison, the IIW FAT-71 class [25] for the used specimen geometry is also presented in the figure. The calculated characteristic fatigue strength (FAT), the mean fatigue strength at 2 million cycles and the slope of the curves are tabulated in Table 4. It is clear that fatigue strength of LTT welds are significantly higher than for welds produced using the conventional wire (Fig. 4 and Table 4). The fatigue strength of the double-pass LTT welds is the highest, followed by the single-pass LTT welds and finally the single and double-pass welds produced by the conventional wire. From the FAT values in Table 4 it can be seen that a fatigue strength improvement of 22% and 55% is achieved for L1 and L2 welds, respectively, when compared with the IIW FAT value of 71 MPa. Experimental FAT-values for welds produced by the conventional filler material were slightly below the IIW FAT 71. Comparing mean fatigue strength levels at 2 million cycles even more clearly shows the higher fatigue strength of the LTT welds. A direct comparison shows an improvement of 58% for the single-pass welds and 76% for the double-pass welds. As can be seen in Fig. 4, the fatigue strength is almost similar for all consumables at high stresses, but higher for LTT consumables at lower loads. The slope of the S–N lines are 3.74 and 4.41 (Table 4) for welds by LTT fillers. These are higher than the slope of 3 used
Table 1 Chemical composition of base material and filler materials (weight %).
Weldox 700 LTT OK Tubrod 14.03a a
Nominal composition.
C
Si
Mn
P
S
Cr
Ni
Mo
V
Ti
Cu
Al
Nb
0.13 0.02 0.07
0.3 0.7 0.6
1.18 1.3 1.7
0.011 0.005 –
0.003 0.007 –
0.27 12.8 –
0.04 6.2 2.3
0.13 0.1 0.6
0.007 – –
0.013 – –
0.01 – –
0.041 – –
0.022 – –
162
E. Harati et al. / International Journal of Fatigue 77 (2015) 160–165 Table 3 Welding parameters. Weld
Filler material
Number of beads
Current (A)
Voltage (V)
Welding speed (mm/s)
Heat input (kJ/mm)
L1 L2 C1 C2
LTT LTT OK Tubrod 14.03 OK Tubrod 14.03
1 2 1 2
285 280 240 250
23.9 23.9 23.5 20.5
6.5 10 5.2 8.8
1.05 0.67 1.09 0.58
Fig. 1. Design and dimensions of the cruciform test specimens (dimensions in mm).
Fig. 2. Photograph of a test specimen.
for the FAT 71 curve and also higher than those of the conventional welds at 2.42 and 2.79. This results in significantly higher fatigue strength for LTT fillers at lower stresses.
Fig. 4. Fatigue test results for cruciform specimens produced using LTT (L1 and L2) and conventional filler materials (C1 and C2). Testing was done with constant amplitude loading at R = 0.1.
3.2. Residual stresses Fig. 5 shows the longitudinal surface residual stresses as function of the distance from the weld toe. The measurements were made in front of weld No. 4 in Fig. 1. In general the measured tensile residual stresses are higher for welds produced with the conventional wire compared to those welded with the LTT wire. The difference is most pronounced very close to the weld toe. For example, the tensile residual stress were almost 200 MPa lower, 1 mm from the weld toe, for single-pass samples welded using LTT compared to single-pass welds using the conventional wire.
Table 4 Characteristic fatigue strength (FAT), mean fatigue strength at 2 million cycles and slope of the S–N curves (m). Weld
FAT (MPa)
Mean fatigue strength (MPa)
m
L1 L2 C1 C2
87 110 65 69
117 137 74 78
3.74 4.41 2.42 2.79
Fig. 3. (a) Weld sample (b) Schematic geometrical features at the weld toe.
163
E. Harati et al. / International Journal of Fatigue 77 (2015) 160–165
Fig. 5. Longitudinal surface residual stresses (stresses transverse to the weld toe) in front of the weld toe for samples welded with LTT and conventional filler materials. Note the lower residual stresses near the weld toe for LTT welds.
It can also be seen that tensile residual stresses are lower at 1 mm compared to 2 mm from the weld toe for LTT welds, whereas the opposite is true for the conventional welds. At distances closer than 3 mm to the weld toe, residual stress is slightly higher for a double-pass LTT weld compared to the single-pass LTT. The trend is opposite at distances further away. For welds made using the conventional wire higher tensile residual stresses were found up to 8 mm from the weld toe for the single-pass samples, compared to double-pass welds. Further away from the weld toe they show almost the same values. When comparing the two welds produced with conventional consumables and the two welds produced with LTT consumables it should be noted that the differences within the two pairs are not the same, close to the weld toe. The difference between the two LTT welds is about 30 MPa. The difference between the conventional welds is about 40–50 MPa. The experimental uncertainty, close to the weld toe, for all welds is in the order of 4– 10 MPa and it reaches to a maximum of about 20 MPa at distances further away. 3.3. Weld toe radius and angle Table 5 shows the results of the weld toe radius and angle measurements using the WIA method. The largest values for both the weld toe radius and angle is found for the single-pass weld produced by conventional filler material (C1). No significant difference can be seen between L1, L2, and C2. The maximum and minimum values for toe radii and angles in Table 5 also show that C1 has the highest values but the ranking between the other welds is uncertain. 4. Discussion The aim of this paper is to compare the relative influence of toe geometry and residual stress on the fatigue life of cruciform welded joints. The results of the testing presented in the results section certainly point to that residual stress has a relatively larger influence than toe geometry. This is based on the fact that a significantly higher mean fatigue strength value was obtained for welds L1 and L2 than for the welds made using conventional filler materials. In particular weld C1, having a significantly larger weld toe radius than the other welds, would be expected to have a higher
mean fatigue strength if the weld toe radius were the most influential parameter. An interesting comparison of the relative effects of stress and toe radius can be made for the two welds produced with the conventional filler material (C1 and C2). These welds show almost identical mean fatigue strength, but C1 has a toe radius of 2.6 mm and C2 a radius of 1.4 mm (Table 5). Assuming the same difference in residual stress at the weld toe as 1 mm away, as results by Ramjaun et al. [30] suggest is reasonable for conventional welding consumables, this makes a quantitative comparison possible. A decrease in residual stress by about 40 MPa (Fig. 5) would then balance the decrease in toe radius from 2.6 mm to 1.4 mm. This result can be compared to those by Björk et. al. [31] who discussed the relationship between the weld toe geometry and fatigue strength for non-load carrying X-joints. From Fig. 15 in their paper a clear trend can be seen that the fatigue strength increases with increasing weld toe radius. Although there is significant scatter in the data, typically a 1 mm increase in the weld toe radius from 1.5 mm to 2.5 mm resulted in about 20–30 MPa increase in FAT. Although speculative this suggests that FAT of weld C1 should have been approximately 24–36 MPa higher than for C2 if residual stresses were similar. In other words it seems that a decrease in residual stress of 40 MPa should correspond to an increase on FAT of about 24–36 MPa in this particular case. This can be compared with results of studies by Nguyen and Wahab [32,33] who developed a mathematical model to investigate the effect of weld geometry and residual stresses on fatigue life of butt welds. The geometry is different from the present study but results suggest that an increase of about 2 mm in toe radius has approximately the same effect on fatigue strength as a decrease in residual stress of about 60 MPa. In other words roughly the same relation as found for the welds in the present study. Another observation is that welds L1 and L2 have almost identical weld toe radius and L2 has slightly higher residual stress, at least close to the weld toe. Still, the mean fatigue strength of L2 is significantly higher than for L1. If the residual stress was the most important factor, then a slightly lower mean fatigue strength would have been expected for L2. This seemingly contradictory result can most likely be explained in terms of the location of residual stress measurement and the technique used. Very recent results obtained by Ramjaun et al. [30] comparing the near surface (0.15 mm below surface) weld toe residual stresses for butt welds produced with conventional and LTT consumables using neutron and X-ray diffraction show interesting results. Very similar stresses were measured by X-ray and neutron diffraction at some distance from the weld toe. However, it was also shown that the X-ray method is not capable of capturing a dramatic local drop in stresses that occur within about 1 mm around the weld toe for LTT welds. For example both X-ray and neutron diffraction measured a transverse stress (normal to the weld line) of about 200 MPa at 1– 5 mm from the weld toe. With neutron diffraction it was also seen that there was a dramatic local change in stress to about 600 MPa at the weld toe fusion boundary. This is the local stress that is more likely to be relevant for fatigue properties than the stress measured at a distance 1 mm from the weld toe. No such change was observed for welds produced with conventional consumables where stresses were fairly constant within a few millimetres of
Table 5 Results from weld toe radius and angle measurements, Average from six measurements together with the maximum and minimum values. Weld
Radius (mm)
Maximum radius (mm)
Minimum radius (mm)
Angle (°)
Maximum angle (°)
Minimum angle (°)
L1 L2 C1 C2
1.4 1.3 2.6 1.4
1.5 1.6 3.1 1.7
1.2 1.0 2.2 1.0
128 134 149 119
145 136 159 140
127 131 148 120
164
E. Harati et al. / International Journal of Fatigue 77 (2015) 160–165
the weld toe. Therefore, X-ray measurement cannot be expected to accurately measure the actual local weld toe stress for LTT welds but can give an indication of whether the residual stresses are higher or lower than for conventional welds. It can also be concluded that they are most likely significantly lower than measured by X-ray at 1 mm. The higher fatigue strength of L2 in comparison to L1 can also be understood in the light of results from the same study by Ramjaun et. al. [30]. They found that the volume of the LTT weld metal has an effect on the local weld toe stress level. For a three-pass LTT weld compressive stresses of about 600 MPa were measured whereas the compressive stresses at a single-pass LTT weld was about 400 MPa. In both cases stresses next to the weld was about 200 MPa with the average stresses somewhat less compressive for the single pass weld. There are obviously geometrical differences between the welds in this paper and the referenced study that may result in different stress levels but it is likely that the overall behaviour is similar. Based on this it is likely that the greater volume of the double-pass weld L2 has contributed to a larger compressive weld toe stress than for the single-pass weld L1 even though X-ray measurements are not capable of resolving this. Furthermore the second weld pass of L2 will have a composition with a lower Ms-temperature due to less dilution and thereby more effectively lowering the residual stress level. Considering the above discussion, the much better fatigue properties of weld L2 than for L1 can therefore most likely be explained in terms of a lower residual stress locally at the weld toe. In conclusion the present results suggest that a moderate change in residual stress of about 15% at the 300 MPa stress level has the same effect on fatigue strength as the increase of the weld toe radius by approximately 85%. It should be noted that this result was obtained for radii ranging from 1.4 mm to 2.6 mm and it cannot be concluded from the present study whether the relation is linear and applies to other weld toe radii or stress ranges or not. The large effect of using LTT consumables than having a larger weld toe radius also shows the importance of the residual stress level. Certainly there is as always significant scatter in the fatigue data, variation in the toe radius along the weld and uncertainties in the measurement of the local stress at the weld toe. Nevertheless the results indicate that a decrease of the residual stress has relatively seen a larger effect on fatigue life than a change in weld toe radius. This is in many way quite reassuring as it is well known, and inevitably so, that there will always be significant variations in the local weld toe radius along the weld as seen in Table 5. Further studies are needed to make a more quantitative comparison and to study whether the effect is the same or varies for different stress levels and toe radii.
For welds made with conventional consumables, a change in residual stress of about 15% at the 300 MPa stress level had the same effect on fatigue strength as the increase of the weld toe radius by approximately 85% from 1.4 mm to 2.6 mm. Residual stress has a relatively larger influence than the weld toe geometry on fatigue strength of cruciform welds.
Acknowledgements The financial support of the Swedish Energy Agency is gratefully acknowledged. We thank ESAB AB for providing the welding filler materials and welding of the samples. The authors would also like to thank Kjell Hurtig (University West, Sweden) for his help with geometrical measurements. References [1] [2] [3]
[4]
[5]
[6]
[7]
[8]
[9]
[10]
[11]
[12]
[13]
[14]
5. Conclusions The fatigue strength of cruciform welds produced with conventional and with LTT filler materials was determined. By comparing welds with different toe radii and residual stress levels the following can be concluded: The fatigue strength of LTT welds at 2 million cycles was 58% higher for single-pass welds and 76% higher for the double-pass welds when compared to corresponding welds made with conventional filler materials. Surface residual stresses very close to the weld toe are higher for welds produced with the conventional wire compared to those welded with the LTT wire. The higher fatigue strength of LTT welds than for welds made with conventional consumables is an effect of the significantly lower stress levels at the weld toe.
[15]
[16] [17]
[18]
[19]
[20]
Stephens RI, Fatemi A, Stephens RR, Fuchs HO. Metal fatigue in engineering. NY: John Willey & Sons; 2001. Pang HLJ. Analysis of weld toe profiles and weld toe cracks. Int J Fatigue 1993;15:31–6. http://dx.doi.org/10.1016/0142-1123(93)90074-Z. Alam MM, Barsoum Z, Jonsén P, Häggblad HA, Kaplan A. Geometrical aspects of the fatigue behaviour of laser hybrid fillet welds. In: Proc. Fatigue Des Conf Senlis Paris Fr; 2009. Atzori B, Lazzarin P, Meneghetti G, Ricotta M. Fatigue design of complex welded structures. Int J Fatigue 2009;31:59–69. http://dx.doi.org/10.1016/ j.ijfatigue.2008.02.013. Kirkhope KJ, Bell R, Caron L, Basu RI, Ma K-T. Weld detail fatigue life improvement techniques. Part 1: Review. Mar Struct 1999;12:447–74. http:// dx.doi.org/10.1016/S0951-8339(99)00013-. Tveiten B, Fjeldstad A, Harkegard G, Myhr O, Bjorneklett B. Fatigue life enhancement of aluminium joints through mechanical and thermal prestressing. Int J Fatigue 2006;28:1667–76. http://dx.doi.org/10.1016/ j.ijfatigue.2006.01.006. Yildirim HC, Marquis GB. Fatigue strength improvement factors for high strength steel welded joints treated by high frequency mechanical impact. Int J Fatigue 2012;44:168–76. Çam G, Özdemir O, Koçak M. Progress in low transformation temperature (ltt) filler wires: review. In: Proc 63rd Annu Assem Int Conf Int Inst Weld, 2010; p. 759–65. Altenkirch J, Gibmeier J, Kromm A, Kannengiesser T, Nitschke-Pagel T, Hofmann M. In situ study of structural integrity of low transformation temperature (LTT)-welds. Mater Sci Eng, A 2011;528:5566–75. http:// dx.doi.org/10.1016/j.msea.2011.03.091. Kromm MA, Kannengiesser T. Characterizing phase transformations of different LTT alloys and their effect on residual stresses and cold cracking. Weld World 2011;55:48–56. http://dx.doi.org/10.1007/BF03321286. Bhatti AA, Barsoum Z, van der Mee V, Kromm A, Kannengiesser T. Fatigue strength improvement of welded structures using new low transformation temperature filler materials. Proc Eng 2013;66:192–201. http://dx.doi.org/ 10.1016/j.proeng.2013.12.07. Takahashi M, Yasuda HY. Variant selection of martensites in steel welded joints with low transformation temperature weld metals. J Alloys Compd 2013;577:S601–4. http://dx.doi.org/10.1016/j.jallcom.2012.02.022. Ooi SW, Garnham JE, Ramjaun TI. Review: low transformation temperature weld filler for tensile residual stress reduction. Mater Des 2014;56:773–81. http://dx.doi.org/10.1016/j.matdes.2013.11.050. Ramjaun TI, Stone HJ, Karlsson L, Kelleher J, Ooi SW, Dalaei K, et al. Effects of dilution and baseplate strength on stress distributions in multipass welds deposited using low transformation temperature filler alloys. Sci Technol Weld Join 2014;19:461–7. http://dx.doi.org/10.1179/1362171814Y.0000000209. Williams HE, Ottsen H, Lawence FV, Munse WH. The effects of weld geometry on the fatigue behavior of welded connrctions. University of Illinois Engineering Experiment Station. College of Engineering. University of Illinois at Urbana-Champaign.; 1970. Niu X, Glinka G. The weld profile effect on stress intensity factors in weldments. Int J Fract 1987;35:3–20. http://dx.doi.org/10.1007/BF00034531. Martins Ferreira JA, Moura Branco CA. Influence of the radius of curvature at the weld toe in the fatigue strength of fillet welded joints. Int J Fatigue 1989;11:29–36. http://dx.doi.org/10.1016/0142-1123(89)90044-3. Teng T-L, Fung C-P, Chang P-H. Effect of weld geometry and residual stresses on fatigue in butt-welded joints. Int J Press Vessels Pip 2002;79:467–82. http://dx.doi.org/10.1016/S0308-0161(02)00060-1. Caccese V, Blomquist PA, Berube KA, Webber SR, Orozco NJ. Effect of weld geometric profile on fatigue life of cruciform welds made by laser/GMAW processes. Mar Struct 2006;19:1–22. http://dx.doi.org/10.1016/ j.marstruc.2006.07.002. Lee C-H, Chang K-H, Jang G-C, Lee C-Y. Effect of weld geometry on the fatigue life of non-load-carrying fillet welded cruciform joints. Eng Fail Anal 2009;16:849–55. http://dx.doi.org/10.1016/j.engfailanal.2008.07.004.
E. Harati et al. / International Journal of Fatigue 77 (2015) 160–165 [21] Radhi HE, Barrans S. Finite Element Analysis of Effect of Weld Toe Radius and Plate Thickness on Fatigue Life of Butt Welded Joint. In: Lucas G, Xu Z, editors. Future Technol. Comput. Eng. Proc. Comput. Eng. Annu. Res. Conf. 2010 CEARC’10, Huddersfield: University of Huddersfield; 2010, p. 60–4. [22] Barsoum Z, Jonsson B. Influence of weld quality on the fatigue strength in seam welds. Eng Fail Anal 2011;18:971–9. http://dx.doi.org/10.1016/j.engfailanal. 2010.12.001. [23] Svensson L-E, Karlsson L, Harati E. Increasing fatigue life using Low Transformation Temperature (LTT) welding consumables. In: Proceeding of 2nd Swedish conference on design and fabrication of welded structures, Borlänge, Sweden; 2013, p. 49–64. [24] Dowling NE. Mechanical behavior of materials : engineering methods for deformation, fracture, and fatigue. Prentice Hall; 1993. [25] Hobbacher A. Recommendations for fatigue design of welded joints and components. International Institute of Welding; XIII-2460-13/XV-1440-13, Paris, France; 2014. [26] Hou C. Fatigue analysis of welded joints with the aid of real three-dimensional weld toe geometry. Int J Fatigue 2007;29:772–85. http://dx.doi.org/10.1016/ j.ijfatigue.2006.06.007. [27] Peng T. Algorithms and models for 3-D shape measurement using digital fringe projections. Doctoral theis: University of Maryland, College Park, USA; 2007.
165
[28] Stenberg T, Lindgren E, Barsoum Z. Development of an algorithm for quality inspection of welded structures. Proc Inst Mech Eng Part B J Eng Manuf 2012;226:1033–41. http://dx.doi.org/10.1177/0954405412439138. [29] Harati E, Ottosson M, Karlsson L, Svensson L-E. Non-destructive measurement of weld toe radius using Weld Impression Analysis, Laser Scanning Profiling and Structured Light Projection methods. In: First Int Conf Weld Non Destr Test Tehran, Iran; 2014. [30] Ramjaun TI, Stone HJ, Karlsson L, Gharghouri MA, Dalaei K, Moat RJ, et al. Surface residual stresses in multipass welds produced using low transformation temperature filler alloys. Sci Technol Weld Join 2014;19:623–30. http://dx.doi.org/10.1179/1362171814Y.0000000234. [31] Björk T, Samuelsson J, Marquis G. The need for a weld quality system for fatigue loaded structures. Weld World 2008;52:34–46. http://dx.doi.org/ 10.1007/BF03266615. [32] Ninh NT, Wahab MA. The effect of residual stresses and weld geometry on the improvement of fatigue life. J Mater Process Technol 1995;48:581–8. http:// dx.doi.org/10.1016/0924-0136(94)01697-Y. [33] Nguyen TN, Wahab MA. The effect of weld geometry and residual stresses on the fatigue of welded joints under combined loading. J Mater Process Technol 1998;77:201–8. http://dx.doi.org/10.1016/S0924-0136(97)00418-4.