Weld formation mechanism of fiber laser oscillating welding of austenitic stainless steel

Weld formation mechanism of fiber laser oscillating welding of austenitic stainless steel

Journal of Materials Processing Technology 225 (2015) 77–83 Contents lists available at ScienceDirect Journal of Materials Processing Technology jou...

2MB Sizes 1 Downloads 47 Views

Journal of Materials Processing Technology 225 (2015) 77–83

Contents lists available at ScienceDirect

Journal of Materials Processing Technology journal homepage: www.elsevier.com/locate/jmatprotec

Weld formation mechanism of fiber laser oscillating welding of austenitic stainless steel Kangda Hao, Geng Li, Ming Gao ∗ , Xiaoyan Zeng Wuhan National Laboratory for Optoelectronics, Huazhong University of Science and Technology, Wuhan 430074, China

a r t i c l e

i n f o

Article history: Received 4 April 2015 Accepted 21 May 2015 Available online 27 May 2015 Keywords: Laser welding Oscillating Morphology Welding mode Stainless steel

a b s t r a c t The effects of beam oscillating parameters on weld morphologies are investigated. Weld geometrical sizes were counted statistically. The results showed that the difference in cross-section width from the top to the lower gradually reduced to disappear with the increase of oscillating frequency. Weld cross-section was featured by slender nail when the frequency was 10 Hz or so, which was similar with conventional laser keyhole weld. It changed to dumpy nail, V-shape and U-shape in sequence with the increase of the frequency within the range from 20 Hz to 1000 Hz. The welding speed thresholds transforming the welding from keyhole mode to unstable mode and from unstable mode to heat conduction mode were confirmed as 24 m/min and 48 m/min, respectively. Weld formation mechanisms were summarized by weld overlapping ratio and the transformation of welding mode. © 2015 Elsevier B.V. All rights reserved.

1. Introduction It is well known that laser welding has the advantages of deep penetration, high aspect ratio, narrow heat-affected zone (HAZ), fine grains and excellent mechanical properties due to high energy density. However, critical edge preparation and resultant clamping accuracy are indispensable for laser welding since laser beam is focused on a very small spot. Moreover, the defects of crack and porosity easily occur in laser welds. Lehner et al. (1999) found that the gap width should be controlled below 10% of workpiece thickness in disk laser welding of AZ91D magnesium alloy because larger gap may cause the defect of weld sag. Katayama (2013) demonstrated that fast solidification rate was one of the main reasons to increase crack within laser welds of aluminum alloy and austenitic stainless steel. Meng et al. (2014) found that the porosity easily formed in CO2 laser lap welding of 4 mmthick low alloy steel because of the rapid solidification of molten pool. In order to improve gap tolerance and weld quality of laser welding, Rubben et al. (1997) developed laser welding with beam oscillating along the weld, which is called laser oscillating welding for brief. By using beam oscillating with maximum frequency of 200 Hz, the gap tolerance of tailed blanks was increased to 0.3 mm, 25% of sheet thickness. With rapid development of gal-

∗ Corresponding author. Tel.: +86 27 87793527; fax: +86 27 87541423. E-mail address: [email protected] (M. Gao). http://dx.doi.org/10.1016/j.jmatprotec.2015.05.021 0924-0136/© 2015 Elsevier B.V. All rights reserved.

vanometer scanner recently, laser oscillating welding has attracted more and more attentions. Schedewy et al. (2008) found that when using high beam oscillating frequency of 100–200 Hz, fiber laser oscillating welding of carbon steel becomes unstable because the liquid metal is thrown out of molten pool by fast moving laser beam. Choi et al. (2010) found that crack susceptibility of 5J32-T4 aluminum alloy welds can be minimized by laser oscillating welding with the frequency of 5 Hz. Kim et al. (2011) showed that the shear-tensile strength of 6k21 aluminum alloy joint made by disk laser oscillating welding is 29% higher than that of conventional laser welding. Vanska and Salminen (2012) claimed that laser oscillating welding could increase the safety factor of industrial applications by widening the weld. Yamazaki et al. (2013) found that increasing beam oscillating frequency widens but shallows the weld. Schweier et al. (2013) demonstrated that either increasing welding speed or decreasing laser power helps reduce the spatter number during laser oscillating welding. Above-mentioned researches indicates that laser oscillating welding is potential to improve industrial adaptability of laser welding by improving gap tolerance, and has its own features in weld morphologies. Relevant studies, especially weld formation mechanisms are few. To deepen the understanding of this new technique, laser oscillating welding of AISI304 stainless steel is studied, and the effects of beam oscillating parameters on weld morphologies are discussed.

78

K. Hao et al. / Journal of Materials Processing Technology 225 (2015) 77–83

Table 2 Welding parameters used in this study. Welding parameters

Value(s)

Laser power, P (kW) Feeding speed of welding head, vx (m/min) Flow rate of argon shielding gas, (L/min) Oscillating amplitude of laser beam, a (mm) Oscillating frequency of laser beam, f (Hz)

2.0 3.0 18 0.5, 1.0, 1.5, 2.5 10, 20, 50, 100, 200, 500, 1000

2. Experimental Base metal used is 3 mm-thick AISI304 austenitic stainless steel with chemical compositions in Table 1. The welding system is composed of a fiber laser source, a galvanometer scanner and a six-axis robot, as shown in Fig. 1. The fiber laser is IPG YLS-6000 with the maximum power of 6 kW. The galvanometer scanner is SCANLAB hurrySCAN30. The six-axis robot is KUKA KR60HA. During welding, the welding head is driven by the robot. The laser beam with the wavelength of 1070 nm is transmitted by a fiber to a collimator with focal length of 200 mm, reflected by a copper mirror to the mirrors within galvanometer scanner, and focused by a lens with focal length of 250 mm to irradiate base metal with a 0.4 mm-diameter beam spot. For brief, as shown in Fig. 1, the feeding direction of welding head is defined as X-axis, and the oscillating direction of laser beam is defined as Y-axis,. The moving speed of welding head is named as feeding speed, vx , and the moving speed of laser beam driven by galvanometer scanner is named as oscillating speed, vy . The vy can be computed as follows.

vy = 4a/T = 4af

(1)

where, T is the time of one cycle, f is beam oscillating frequency, a is beam oscillating amplitude. The real speed of laser beam resulted by vx and vy is denoted as welding speed, v, and the real moving direction of laser beam is named as welding direction. The v can be computed as follows.



v=

v2x + v2y

(2)

The welding parameters used are listed in Table 2. The experiments were carried out in bead-on-plate configuration. After welding, the welds were cut down to prepare the metallurgical samples. The samples were etched by a solution of 8 g CuCl2 , 100 mL HCl and 100 mL C2 H5 OH. Both surface and cross-section morphologies were examined and measured by optical microscope. 3. Results 3.1. Weld surface morphology As shown in Fig. 2, the welds begin to overlap from the edge to the center with the increase of oscillating frequency. Under each oscillating amplitude, weld surface is characterized by zigzag line, sawtooth line and smooth line in sequence with the increase of the frequency, indicating the fraction of overlapping zone is increasing. Taking the welds with amplitude of 0.5 mm as example, the weld surface is presented as zigzag line when the frequency is 10 Hz, and then changes to be sawtooth line when the frequency reaches 50 Hz, and finally changes to be smooth line after the frequency is larger than 200 Hz. 3.2. Weld cross-section morphology As shown in Fig. 3, increasing either beam oscillating frequency or amplitude decreases weld penetration depth, and reduces the difference between the cross-section width at the surface, cws and

the cross-section width at half depth, cwm . Four types of weld cross-section morphologies appear in sequence with the increase of oscillating frequency, which are slender nail, dumpy nail, V-shape and U-shape. According to the shape characteristics of conventional laser welds, the nail shape, V-shape and U-shape are corresponding to keyhole mode, unstable mode between keyhole mode and heat conduction mode, and heat conduction mode, respectively. Along the transverse width, moreover, the penetration depth at the edge of V-shape and U-shape welds is bigger than that at the center since the edges are heated repeatedly by reversed laser beam. 3.3. Weld geometrical characteristics The geometrical characteristics of weld surface are illustrated in Fig. 4. For brief, the widths of overlapping and non-overlapping zones along Y-axis are named as wo and wno , respectively. The whole width of weld surface is named as w. The w is larger than theoretical value, double amplitudes since the molten pool itself has a width. The extra increment caused by molten pool is named as e. Thus, the w can be expressed as follows. w = 2(wo + wno ) = 2(a + e)

(3)

The fraction of overlapping zone to whole width of weld surface is defined as overlapping ratio, ı, which can be expressed by Eq. (4). ı = 2wo /w

(4)

In Fig. 5, the w decreases gradually to equal to or even smaller than double amplitudes with the increase of oscillating frequency. Schweier et al. (2011) claimed that galvanometer scanner has a frequency limit related to given amplitude. Once the frequency exceeds the limit, the scanner inertia will prevent the laser beam reaching reversal points, which reduces the real amplitude to be smaller than setting value. Meanwhile, the molten pool at this time is too narrow to compensate the reduction because high oscillating frequency causes a too fast welding speed. As a result, the w is even narrower than double amplitudes when the frequency is large enough. Since this frequency limit increases with the decrease of oscillating amplitude, the frequency threshold reducing the w to be smaller than double amplitudes varies with the amplitude. For given amplitudes of 1.0 mm, 1.5 mm and 2.5 mm, the frequency thresholds are found as 800 Hz, 500 Hz and 100 Hz, respectively. However, this phenomenon disappears when the amplitude is 0.5 mm because its frequency limit is higher than 1000 Hz. In addition, a strong relationship between the ı and crosssection widths is found in Fig. 5. Firstly, both the difference between cws and cwm and the difference between w and cws decrease with the increase of the ı. It implies that the weld is more homogeneous in cross-section width. Secondly, the cws is nearly equal to the w when the weld overlaps completely. The welds with the amplitude of 0.5 mm overlap completely when the frequency is higher than 200 Hz, while the welds with the amplitude equal to or larger than 1.0 mm overlap completely when the frequency is larger than 500 Hz. In Fig. 6a, weld penetration depth decreases sharply as the frequency increases from 10 Hz to 100 Hz or so, and then drops slowly, but nearly keeps stable after the frequency reaches 200 Hz. It would be caused by the variation of welding mode. When the frequency is smaller than 100 Hz, the heat input is sufficient to maintain keyhole mode. Thus, the penetration depth is inversely proportional to the frequency since the welding speed is proportional to the frequency. When the frequency increases to be larger than 200 Hz, the welding is characterized by heat conduction mode because the heat input becomes too small to keep keyhole mode. As a result, the penetration at this stage is shallow and nearly without fluctuations.

K. Hao et al. / Journal of Materials Processing Technology 225 (2015) 77–83

79

Fig. 1. Schematic drawing of experimental set-up, where vx is feeding speed of welding head, vy is oscillating speed of laser beam, and v is real speed of laser beam.

Fig. 2. Weld surface morphologies under different oscillating parameters. (a) a = 0.5 mm, f = 10 Hz; (b) a = 0.5 mm, f = 50 Hz; (c) a = 0.5 mm, f = 100 Hz; (d) a = 0.5 mm, f = 200 Hz; (e) a = 1.0 mm, f = 10 Hz; (f) a = 1.0 mm, f = 50 Hz; (g) a = 1.0 mm, f = 100 Hz; (h) a = 1.0 mm, f = 500 Hz; (i) a = 1.5 mm, f = 10 Hz; (j) a = 1.5 mm, f = 50 Hz; (k) a = 1.5 mm, f = 100 Hz; (l) a = 1.5 mm, f = 500 Hz; (m) a = 2.5 mm, f = 10 Hz; (n) a = 2.5 mm, f = 50 Hz; (o) a = 2.5 mm, f = 100 Hz; (p) a = 2.5 mm, f = 500 Hz.

80

K. Hao et al. / Journal of Materials Processing Technology 225 (2015) 77–83

Table 1 Chemical compositions of AISI304 austenitic stainless steel (wt.%). Element

C

Mn

P

S

Si

Cr

Ni

Fe

wt.%

0.08

2.0

0.045

0.03

1.0

18.0–20.0

8.0–10.5

Balance

Fig. 3. Effects of oscillating parameters on weld cross-section morphologies, where d is penetration depth, cws is cross-section width at the surface, and cwm is cross-section width at half depth. (a) a = 0.5 mm, f = 10 Hz; (b) a = 0.5 mm, f = 100 Hz; (c) a = 1.0 mm, f = 10 Hz; (d) a = 1.0 mm, f = 50 Hz; (e) a = 1.0 mm, f = 100 Hz; (f) a = 1.0 mm, f = 200 Hz; (g) a = 1.5 mm, f = 10 Hz; (h) a = 1.5 mm, f = 50 Hz; (i) a = 1.5 mm, f = 100 Hz; (j) a = 1.5 mm, f = 200 Hz; (k) a = 2.5 mm, f = 10 Hz; (l) a = 2.5 mm, f = 20 Hz; (m) a = 2.5 mm, f = 50 Hz; (n) a = 2.5 mm, f = 100 Hz.

Fig. 4. Schematic drawing of weld surface characteristics, where w is whole width of weld surface, wo is the width of overlapping zone along Y-axis, wno is the width of non-overlapping zone along Y-axis, a is beam oscillating amplitude, e is extra increment of the width caused by molten pool.

K. Hao et al. / Journal of Materials Processing Technology 225 (2015) 77–83

81

Fig. 5. Weld widths and overlapping ratio, where w is the whole width of weld surface, cws is the cross-section width at surface, cwm is the cross-section width at half depth, and ı is overlapping ratio. (a) a = 0.5 mm, (b) a = 1.0 mm, (c) a = 1.5 mm, (d) a = 2.5 mm.

a=0.5mm a=1.0mm a=1.5mm a=2.5mm

(a)

2.5 2.0 1.5 1.0 0.5 0.0 0

200

400

600

800

(b)

Weld penetration depth, (mm)

Weld penetration depth, (mm)

3.0

keyhole mode unstable mode heat conduction mode

2.5 2.0

v=24 m/min

1.5

v=48 m/min

1.0 0.5 0.0

1000

Beam oscillating frequency, (Hz)

0

100

200

300

400

500

600

Welding speed, (m/min)

Fig. 6. Weld penetration depth as a function of beam oscillating frequency (a) and welding speed (b).

Table 3 Calculated frequency thresholds to transform welding mode. Oscillating amplitude of laser beam, mm

0.5

1.0

1.5

2.5

Frequency threshold to transform the welding from keyhole mode to unstable mode, Hz Frequency threshold to transform the welding from unstable mode to heat conducted mode, Hz

198.4 399.2

99.2 199.6

66.1 133.1

39.7 79.8

82

K. Hao et al. / Journal of Materials Processing Technology 225 (2015) 77–83

Fig. 7. Schematic diagram of weld cross-section formation: (a) slender nail under keyhole mode, (b) dumpy nail under keyhole mode, (c) V-shape under unstable mode, (d) U-shape under heat conduction mode.

Since both beam oscillating frequency and amplitude are closely associated to the welding speed, the relationship between weld penetration depth and welding speed is illustrated in Fig. 6b to describe the variation of welding mode quantitatively. The relationship can be fitted as Eq. (5).

d=

⎧ ⎪ ⎨ 2.617 − 0.053v(0 < v ≤ 24m/min)

380.420 × v−1.774 (24m/ min ≤ v ≤ 48m/min) }

⎪ ⎩ 0.518 − 6.35 × 10−4 v(v > 48m/min)

(5)

The equation shows that the welds with welding speed lower than 24 m/min are under keyhole mode, whose penetration depth is inversely proportional to and decreases sharply with welding speed. The welds with welding speed larger than 48 m/min are under heat conduction mode, whose penetration depth is inversely proportional to but decreases slowly with welding speed. The welds with welding speed ranging from 24 m/min to 48 m/min are under unstable mode, whose penetration depth decreases with the power of welding speed. According to Fig. 6b and Eq. (5), the frequency thresholds to transform the welding mode from keyhole mode to unstable mode and from unstable mode to heat conducted mode are calculated in Table 3.

4. Discussion Assuming that laser beam initially moves from the center to the edge at negative Y-axis after a quarter cycle, reverses and passes through the center to the edge at positive Y-axis after three quarter cycle, and returns to the center after one cycle, the formation of weld cross-section shape can be illustrated in Fig. 7. Where, the center and the reversal point at positive Y-axis are named as position A and position B, respectively. When the frequency is 10 Hz or so, the time interval from position A to position B is too long due to a slow welding speed. The earlier formed molten pool at position A has solidified when laser beam arrives at position B, as shown in Fig. 7a. It denotes that the beam oscillating has no effect on the welding mode at this stage.

Thus, the weld is represented as a slender nail, similar with conventional laser weld. When the frequency is lower than the frequency threshold changing the welding mode from keyhole to unstable, the welding still keeps keyhole mode with the increase of the frequency, although the welding speed increases rapidly. Fabbro (2010) claimed that the molten pool is elongated and narrowed by increasing welding speed. Then, the rapidly increased welding speed makes the molten pool at position A keep liquid when laser beam moves to position B, as shown in Fig. 7b. Obviously, the molten pools of position A and position B combine partly, which widens the lower part of final molten pool and causes a dump nail weld. When the frequency is between the two thresholds in Table 3, the welding fluctuates between keyhole mode and heat conduction mode. The molten pool is composed of slender molten pool caused by keyhole mode and shallow arc-shape molten pool caused by heat conduction mode, as shown in Fig. 7c. Consequently, the weld cross-section represents a V-shape at this stage. When the frequency increases to be larger than the frequency threshold changing the welding from unstable mode to heat conduction mode, there is only the overlapping of shallow arc-shape molten pools, as shown in Fig. 7d. The weld also overlaps completely at this stage. It then results in a weld cross-section featured by U-shape with the width of double amplitudes. 5. Conclusions The effects of beam oscillating parameters on the morphologies of fiber laser welded AISI304 stainless steel are investigated in detail. The conclusions are drawn as follows. 1. The overlapping ratio of the weld increases gradually with the increase of oscillating frequency. It results in that, under each given amplitude, weld surface morphology varies from zigzag line to sawtooth line and smooth line in sequence with the increase of the frequency. 2. Weld cross-section becomes more homogeneous with the increase of oscillating frequency because the difference in cross-

K. Hao et al. / Journal of Materials Processing Technology 225 (2015) 77–83

section width from the top and the lower gradually reduces to disappear. With the increase of the frequency, the weld crosssection changes from slender nail to dumpy nail, V-shape and U-shape in sequence. 3. The welding mode is closely related with welding speed. The welding can be divided into three modes with the increase of welding speed, which are keyhole mode while welding speed lower than 24 m/min, unstable mode while welding speed between 24 m/min and 48 m/min, and heat conduction mode while welding speed larger than 48 m/min. The weld penetration depth is inversely proportional to and decreases sharply with welding speed under keyhole mode, while it decreases slowly under unstable mode and keeps nearly stable under heat conduction mode. 4. The effects of oscillating frequency on weld cross-section are dependent on weld overlapping ratio and the variation of welding mode. Acknowledgements This research is financially supported by the National Natural Science Foundation of China (Grant nos. 51275186, and 51475183), and the Fundamental Research Funds for the Central Universities of China, HUST. References Choi, K.D., Ahn, Y.N., Kim, C., 2010. Weld strength improvement for Al alloy by using laser weaving method. J. Laser. Appl. 22 (3), 116–119.

83

Fabbro, R., 2010. Melt pool and keyhole behavior analysis for deep penetration laser welding. J. Phys. D: Appl. Phys. 43, 445501. Katayama, S., 2013. Handbook of Laser Welding Technologies. Woodhead Publishing, Cambridge, pp. 341–345. Kim, B.H., Kang, N.H., Oh, W.T., Kim, C.H., Kim, J.H., Kim, Y.S., Park, Y.H., 2011. Effects of weaving laser on weld microstructure and crack for Al 6k21-T4 alloy. J. Mater. Sci. Technol. 27 (1), 93–96. Lehner, C., Reinhart, G., Schaller, L., 1999. Welding of die cast magnesium alloys for production. J. Laser. Appl. 11 (5), 206–210. Meng, W., Li, Z.G., Lu, F.G., Wu, Y.X., Chen, J.H., Katayama, S., 2014. Porosity formation mechanism and its prevention in laser lap welding for T-joints. J. Mater. Process. Technol. 214, 1658–1664. Rubben, K., Mohrbacher, H., Leirman, E., 1997. Advantages of using an oscillating laser beam for the production of tailed blanks. SPIE, 3097. 228-241. Schedewy, R., Dittrich, D., Standfuß, J., Brenner, B., 2008. Lbw of high stiffness light weight structures generated by scanned fiber laser beams. In: Proceedings of the International Congress on Application of Lasers & Electro-Optics (ICALEO), Temecula, USA, pp. 332–339. Schweier, M., Hatwig, J., Zaeh, M.F., Reppich, J., 2011. Single mode fiber laser beam welding with superposed beam oscillating. In: Proceedings of the International Congress on Application of Lasers & Electro-Optics (ICALEO), Florida, USA, pp. 536–546. Schweier, M., Heins, J.F., Haubold, M.W., Zaeh, M.F., 2013. Spatter formation in laser welding with beam oscillation. Phys. Procedia. 41, 20–30. Vanska, M., Salminen, A., 2012. Laser welding of stainless steel self-steering tube-to-tube joints with oscillation mirror. J. Eng. Manuf. 226, 632–640. Yamazaki, Y., Abe, Y., Kitagawa, A., Nakata, K., 2013. Fundamental Study on the Laser Welding Phenomena with High-frequency Laser Beam Oscillation. In: Proceedings of LAMP2013-the 6th International Congress on Laser Advanced Materials Processing, Niigata, Japan, pp. 1–7.