Effects of beam configurations on wire melting and transfer behaviors in dual beam laser welding with filler wire

Effects of beam configurations on wire melting and transfer behaviors in dual beam laser welding with filler wire

Optics & Laser Technology 91 (2017) 138–148 Contents lists available at ScienceDirect Optics & Laser Technology journal homepage: www.elsevier.com/l...

2MB Sizes 1 Downloads 59 Views

Optics & Laser Technology 91 (2017) 138–148

Contents lists available at ScienceDirect

Optics & Laser Technology journal homepage: www.elsevier.com/locate/optlastec

Full length article

Effects of beam configurations on wire melting and transfer behaviors in dual beam laser welding with filler wire

MARK



Guolong Ma, Liqun Li , Yanbin Chen State Key Laboratory of Advanced Welding and Joining, Harbin Institute of Technology, Harbin 150001, China

A R T I C L E I N F O

A BS T RAC T

Keywords: Dual beam Laser welding Filler wire Beam configuration Transfer behavior

Butt joints of 2 mm thick stainless steel with 0.5 mm gap were fabricated by dual beam laser welding with filler wire technique. The wire melting and transfer behaviors with different beam configurations were investigated detailedly in a stable liquid bridge mode and an unstable droplet mode. A high speed video system assisted by a high pulse diode laser as an illumination source was utilized to record the process in real time. The difference of welding stability between single and dual beam laser welding with filler wire was also compartively studied. In liquid bridge transfer mode, the results indicated that the transfer process and welding stability were disturbed in the form of “broken-reformed” liquid bridge in tandem configuration, while improved by stabilizing the molten pool dynamics with a proper fluid pattern in side-by-side configuration, compared to sigle beam laser welding with filler wire. The droplet transfer period and critical radius were studied in droplet transfer mode. The transfer stability of side-by-side configuration with the minium transfer period and critical droplet size was better than the other two configurations. This was attributed to that the action direction and good stability of the resultant force which were beneficial to transfer process in this case. The side-by-side configuration showed obvious superiority on improving welding stability in both transfer modes. An acceptable weld bead was successfully generated even in undesirable droplet transfer mode under the present conditions.

1. Introduction Laser welding, as a promising joining technique, has gained considerable acceptance in many industries especially for tailor welded blanks (TWBs) in automotive industry due to its increased productivity and good flexibility [1]. In the production of TWBs, steel sheets are typically welded in butt joint configuration. However, owing to the small focal spot diameter of laser beam, it needs precise fit-up requirement to guarantee the weld quality. Moreover, the gap-bridging capability, which is an important factor on influencing the weld quality and productivity, is also significantly restricted. The largest accepted gap in butt joint is usually 10% of sheet thickness for traditional single beam laser welding, which is often difficult to reach. Previous studies indicated that two effective welding methods, which were dual beam laser welding (DBLW) and single beam laser welding with filler wire (LWFW), could improve the gap-bridging capability obviously [2–5]. Related studies indicated that the use of DBLW could improve gapbridging capability effectively compared to conventional single beam laser welding. Longfield et al. [2] employed this technique to join 2.5 mm thick butt joint and reported that the gap tolerance increased from 0.2 to 0.4 mm compared to single beam laser welding; this



technique was also successfully applied to the industrial production of steel coil joining. For DBLW, the arrangement, interbeam spacing, and energy distribution of dual laser beams can be adjusted flexibly; thereby, a proper beam configuration can be designed for diversiform butt joints with gaps. This characteristic of DBLW results in the increased gap tolerance. In adition to this, DBLW can also improve weld quality effectively in the form of improving weld appearance [6,7], reducing porosity [8], and inhibiting cracking susceptibility [9]. LWFW is another effective way to improve gap-bridging capability [3–5]. The use of additional filler wire provides enough material to fill the gaps and thereby leads to a higher gap tolerance than that of without filler wire. A 2 mm carbon steel thick butt joint with reached up to 1 mm wide air gap was successfully welded by using a CO2 laser with filler wire [5]. The above mentioned results indicate that both DBLW and LWFW can improve the gap-bridging capability of single beam laser welding effectively. Therefore, a dual beam laser welding with filler wire (DBFW) technique is logically proposed to promote the gap-bridging capability further. But only few publications have mentioned the technique and study its gap-bridging capability. Aalderink et al. [10] compared the gap-bridging capability of different laser welding tech-

Correspondence to: Postal address: 92 Xidazhi Street, Nangang District, Harbin City, Heilongjiang Province 150001, PR China. E-mail address: [email protected] (L. Li).

http://dx.doi.org/10.1016/j.optlastec.2016.12.019 Received 2 April 2016; Received in revised form 18 December 2016; Accepted 20 December 2016 0030-3992/ © 2016 Elsevier Ltd. All rights reserved.

Optics & Laser Technology 91 (2017) 138–148

G. Ma et al.

Table 1 Chemical compositions of base metal and filler wire (in wt%).

321 SS ER321

C

Mn

Si

Cr

Ni

Mo

Cu

P

S

Ti

Fe

0.045 0.049

1.08 1.47

0.47 0.50

17.02 19.05

9.02 9.10

– 0.46

– 0.15

0.034 0.024

0.0065 0.020

0.22 0.50

Bal. Bal.

compartively studied. The research findings provided more clear understanding for improving process stability and weld quality.

niques through the welding of 1.1 mm thick aluminum sheets butt joints. The result showed that the gap tolerance increased to 0.6 mm further by using twin-spot laser welding with cold wire feeding, compared to that of 0.4 mm for LWFW process. It definitely proved that the DBFW technique could combine the superiority of these two techniques in gap-bridging capability. For laser welding with filler wire process, there exist numerous technical parameters about the positioning of filler wire, so it is difficult to guarantee the process stability during welding. Most studies [11–14] have focused on the influence of welding parameters on weld quality. Salminen et al. [11–13] reported that wire feeding position and feeding angle were two important factors affecting the process stability, and inaccurate positioning of filler wire could increase the reflectivity of laser and deteriorate the weld quality. Syed [14] studied the effects of wire feeding direction and location on weld quality and found that wire feeding in the front direction and at the leading edge of the melt pool preferred to obtain a better weld quality. For DBFW technique, the welding process becomes more flexible because of the additional welding parameters of dual beam configuration, which is a very important factor. In order to study the influence of wire feeding position on weld quality, rsearchers started to pay attention to the wire melting dynamics which was closely related to the process stability and weld quality [15–17]. High speed imaging technique is very useful to observe the wire melting and transfer behavior. Based on this technique, Yu et al. [15] revealed that the melting dynamics of LWFW were characterized as explosion, big droplet and molten metal bridge according to the distance between the filler wire and laser beam in the welding direction. It was concluded that molten metal bridge could guarantee a stable welding process and good weld quality, while big droplet transition led to disturbed stability and unacceptable weld quality. Tao et al. [16] studied the wire melting behavior and process stability in fiber laser welding of T-joints and similar results were obtained. It was reported that liquid bridge transfer mode was superior to droplet transfer mode with more stable welding process which resulted in better weld appearance and lower porosity defect. Takahashi et al. [17] studied the filler wire meltig dynamics during CO2 laser welding of aluminum alloys using a high-speed shadowgraph imaging technique and also found that humping defect were formed with droplet transfer. Based on the above analyse, it was indicated that there were mainly two wire transfer modes in LWFW process: liquid bridge mode and droplet mode. The liquid bridge mode could lead to a stable welding process and good weld quality, while the droplet mode was inherently unstable and resulted in undesirable weld beads. However, in the practical production, this transfer mode is unavoidable due to restriction of assembly accuracy. So how to inhibit the welding unstability in droplet transfer mode also needs to be studied. Although some researchers had applied the high speed imaging technique to observe the melting wire transfer behaviors, the captured images were still not clear enough to describe the whole process clearly and detailedly. Besides that, to the best of our current knowledge, no results of wire melting and transfer behaviors in DBFW technique have been reported. In this work, a 2 mm thick stainless steel butt joint with 0.5 mm gap was welded by DBFW technique. The objective of this work was to investigate the effects of dual beam configurations on wire melting and transfer behaviors in different transfer modes. Meanwhile, the difference of melting dynamics between DBFW and LWFW was also

2. Materials and experimental procedure 2.1. Material 2 mm thick 321 stainless steel (SS) sheets with the dimension of 150 mm (length)×75 mm (width) were used in this study. The commerical ER321 stainless steel filler wire with the diameter of 1.2 mm was utilized. Their chemical compositions are listed in Table 1. A butt joint with 0.5 mm air gap was adopted in all experiments. Before welding, the surface of stainless steel sheets were polished by abrasive paper for removing the oxides, and then wiped with industrial alcohol. 2.2. Experimental setup Fig. 1 shows the experimental setup of DBFW including a dual beam laser welding system, a wire feeding system and a high speed image acquisition system. The dual beam laser welding system mainly consists of an YW50 welding head based on modular design. The dual laser beams were achieved by a beam splitting module positioned in the welding head, which were then arranged in tandem configuration or side-by-side configuration (as shown in Fig. 1). The interbeam spacing L was kept constant at 0.6 mm. The welding experiments were performed using a continuous wave solid-state Ytterbium fiber laser (YLS-10000, IPG Laser GmbH) with a maximum power of 10 kW. The wavelength of fiber laser is 1.07 µm. A focusing lens of 200 mm and a collimating lens of 150 mm were used to focus the laser beam. The spot diameter of single laser beam at the focal length is 0.26 mm, and the focusing plane was positioned at the surface of workpiece. The spot diameters of the split dual laser beams were treated as the same as that of single laser beam because they used the same collimating lens and focusing lens. In the experiment, the total laser power of the single beam was 1500 W,

Fig. 1. Schematic of experimental setup for DBFW.

139

Optics & Laser Technology 91 (2017) 138–148

G. Ma et al.

Fig. 3. Definition of wire feeding position in DBFW with tandem configuration (side view).

location parameter and its value is colse related to the wire transfer mode. The schematic of wire feeding position is shown in Fig. 3. In the present work, two typical wire transfer modes were chosen as the research objects, which were liquid bridge transfer mode and droplet transfer mode. The effects of dual beam configurations on the wire melting and transfer behaviors in different transfer modes were studied detailedly. Different wire transfer modes were accurately obtained by combining the previous experiment results and related data in literatures and Wx was set in two different values: 0 mm for liquid bridge mode and −1.5 mm for droplet mode. The welding parameters used in DBFW are shown in Table 2.

Fig. 2. Schematic of power density distributions of laser heat sources: (a) single beam and (b) dual beam.

and each split beam had a laser power of 750 W for dual laser beams. The theoretical power density distribution of laser heat source is shown in Fig. 2. The dual laser beams are completely separated with interbeam spacing 0.6 mm and the power density of each laser beam is about one half that of sigle beam with the same power. A wire feeder (KD-4010, Fronius International GmbH) was used to feed the wire. The filler wire was in the leading direction with a filling angle αw 30° with respect to the horizontal surface of the workpiece during welding process. Ar as the shielding gas was supplied to the surface and back of molten pool and wire melting tips with a flow rate of 20 L/min. To stabilize the welding process, the filler wire and shielding gas were set on the same plane with the centerline of dual laser beams. The high speed image acquisition system was composed of a high speed video camera (Camrecord 5000×2, Optronis GmbH) and a high pulse diode laser source (CAVILUX HF, Cavitar Ltd). The high speed video camera was used to observe the wire melting and transfer behaviors and molten pool dynamics during the welding process in real time, which was performed with an angle θ 75° with respect to the horizontal surface of the workpiece, as shown in Fig. 1. The images were recorded at a frame rate of 4000 frams per sec. The high pulse diode laser source which has a wavelength of 810 nm, maximum pulse duration 10 μs and maximum output power 500 W, was used as assistant light source to illuminate the welding zone for obtaining clear images of welding process. To remove the interference of plasma on the acquisited images, a narrow-band pass filter with a wavelength of 810 nm was also attached at the front of high speed camera.

2.4. Weld formation analysis After welding, the weld appearances were observed by an optical microscopy (OLYMPUS SZX 12) firstly. All welds were then sectioned, mounted, polished and etched with 10% mass fraction of Chromium trioxide solution using electrolytic corrosion method at room temperature. The weld cross sections were observed by an optical microscopy (OLYMPUS GX 71). 3. Results and discussion 3.1. Effects of beam configurations in liquid bridge mode Fig. 4 shows the wire melting and transfer behavior in liquid bridge mode when dual beams set in tandem configuration. As shown in Fig. 4(a), one laser beam acted on the molten pool and formed a keyhole; the other beam heated and melted the wire tip, and then a liquid bridge was formed between wire tip and molten pool. According to the contour of wire tip, the bottom of wire tip was still not melted completely, which proved that the laser beam of melting the wire was hardly applied to the molten pool. The melted filler wire then flew Table 2 Welding parameters used in DBFW. Welding parameters

Values

Total laser power Dual beam energy ratio Interbeam spacing (L) Beam configuration Welding speed Wire feeding rate Wire feeding angle (αw) Wire feeding direction Shielding gas flow rate

1500 W 50:50 (750 W+750 W) 0.6 mm Tandem/ side-by-side/single 1 m/min 1.2 m/min 30° leading 20 L/min

2.3. Definition of wire feeding position and transfer mode According to previous studies [15], the relative distance (Wx) between filler wire and laser beam in the welding direction affected the wire melting dynamics significantly. For the tandem configuration, the wire feeding position in DBFW can be defined by relative distance between wire-specimen contact position and the adjacent laser beam, while between wire-specimen contact position and the centerline of dual laser beams for side-by-side configuration. Wx is an important 140

Optics & Laser Technology 91 (2017) 138–148

G. Ma et al.

Fig. 4. Wire melting and transfer behaviors in liquid bridge transfer mode during DBFW with tandem configuration.

downward smoothly into the leading edge of molten pool. This wire transfer mode was defined as liquid bridge transfer mode. However, the liquid bridge was not stable and could be broken in the welding process, as shown in Fig. 4(b). Until accumulating enough amount of liquid metal at the wire tip and connecting with molten pool, a new liquid bridge was reformed, while the keyhole stability was also disturbed, as shown in Fig. 4(c, d). The “broken-reformed” liquid bridge disturbed the welding stability obviously. It should be pointed out that this process was not periodic and occurred randomly. Fig. 5 shows the wire melting and transfer behaviors in liquid bridge mode when dual laser beams set in side-by-side configuration. In this case, the dual laser beams heated the both sides of wire tip simultaneously, and two separated keyholes and liquid bridges were formed. It meant that the energy of both laser beams was applied on wire tip and molten pool. The contour of the wire tip further proved that the wire melting processes of dual beams were divided and no interactions occurred between the two liquid bridges. This assured that both the wire melting and transfer processes were stable in side-by-side configuration of DBFW. Typical wire melting and transfer behaviors in traditional single beam laser welding with filler wire in liquid bridge mode is shown in Fig. 6. A clear high speed image was obtained in the present work. Compared to dual beam laser welding, the power density of single laser

beam was doubled. Therefore, the laser beam heated the wire tip and formed a liquid bridge firstly during the welding process, and then continued to act on the molten pool, in which a bigger but unstable keyhole was established. The fact that the bottom of wire tip was over melted also suggested the laser heating process. The molten pool dynamics was also disturbed and violent fluctuations occurred. The differences of wire transfer behaviors and molten pool dynamics would be reflected on the final weld quality. Fig. 7 shows the weld appearances and cross sections with different beam configurations in liquid bridge mode. For the cross sections, it could be seen that all welds were generally complete without weld defects such as under fill, under cut, lack of fusion and penetration. But the weld appearances were distinguishing, which were influenced by the stability of wire transfer and molten pool. For tandem configuration, the weld appearance had a relativly even surface and rough boundary with some protuberances. For sigle beam, the weld boundary was relatively smooth, but the surfac was rough and irregular because of the unstable keyhole. A few spatters could be found in this case. Among these welds, the weld formation of side-by-side configuration was superior to the others with good weld appearance and symmetric cross section. It should be noted that there existed mismatch in the weld of single beam as shown in Fig. 7(c). This was probably due to the assembly error during preparing butt joint with gap. But in this situation, it was

Fig. 5. Wire melting and transfer behaviors in liquid bridge transfer mode during DBFW with side-by-side configuration.

Fig. 6. Wire melting and transfer behaviors in liquid bridge transfer mode during LWFW.

141

Optics & Laser Technology 91 (2017) 138–148

G. Ma et al.

Fig. 7. Weld formation with different beam configurations in liquid bridge mode: (a) tandem configuration, (b) side-by-side configuration, and (c) single beam.

with small impact on the keyhole. But in the liquid bridge reforming process, an abrupt addition of liquid metal into molten pool would disturbe the keyhole and molten pool stability. The protuberances occurred in the boundary were attributed to this phenomenon as shown in Fig. 7(a). Fluid flow in the molten pool was also important factor influencing welding stability and weld quality. Zhou et al. [18] studied the fluid flow in dual beam laser welding and found that the interaction of two keyholes caused the strong fluid flow in the direction perpendicular to the connection line of the two laser beam centers. The fluid flow pattern of this case was similar to their results. The melted wire flew downward to the molten pool [19], while the liquid metal near the keyhole flew outside of keyhole in the molten pool. The two fluid flows impacted at the front of molten pool and then flew to the direction of molten pool width, as shown in Fig. 8 (b). This fluid flow pattern hit the boundary of molten pool and easily led to rough boundary in the weld appearance.The result indicated that the additional laser beam inducing metal vapor ejecting force had a negative effect on the welding stability in tandem configuration of DBFW. For side-by-side configuration, the wire melting and transfer behaviors had been described in detail and the entire process was stable. The melted wire metal flew downward to the molten pool smoothly in two separated liquid bridges without mutual disturbance. With the influence of the two keyholes, the two fluid flows flew to the rear part of molten pool, as shown in Fig. 8(c). The keyhole and molten pool dynamics were hardly disturbed and relatively stable in this case. Consequently, a good weld formation was obtained. It indicated that the side-by-side configuration could improve the welding stability effectively. For LWFW process, the wire melting and transfer process were still

considered that the mismatch had little influence on the wire melting and transfer behaviors according to the high speed video observation results. 3.2. Analysis of welding stablity in liquid bridge mode In laser beam welding with filler wire, there are mainly three processes affecting the welding stability and weld quality: the wire melting and transfer behavior, the keyhole and molten pool dynamics, and the fluid flow in molten pool. Compared to LWFW, these processes were different due to the additional laser beam acting on the filler wire tip or the molten pool for DBFW. This would introduce additional acting force or change melt flow pattern, which consequently affected the process stability and weld quality. The action mechanisms of wire melting and transfer behaviors on weld quality were analyzed through the above metnioned three aspects, and the results are summarized in Fig. 8. Fig. 8(a) shows the liquid transfer process with tandem configuration. During the welding process, the additional laser beam heated the molten pool and a keyhole was generated accompanied by ejecting drastic metal vapor. The metal vapor would influence the stability of liquid bridge in two aspects. On one hand, a metal vapor ejecting force was applied to the liquid bridge directly and disturbed the liquid bridge transfer process. On the other hand, the front of molten pool was also disturbed where the liquid bridge connected the molten pool. Moreover, for butt joint with gap configuration, the front of molten pool was the gap zone, which resulted in a more unstable molten pool front and was disturbed more easily. Therefore, a “broken-reformed” liquid bridge occurred in the transfer process, as shown in Fig. 4. Normally, the melted wire metal would flow smoothly into molten pool 142

Optics & Laser Technology 91 (2017) 138–148

G. Ma et al.

Fig. 8. Schematic of action mechanisms of wire melting and transfer behaviors with different beam configurations in liquid bridge mode: (a) liquid bridge transfer mode (side view), (b) tandem configuration, (c) side-by-side configuration, and (d) single beam (top view).

In addition, the molten metal would flow to the outside of the keyhole, which was useless to strengthen the welding stability. Therefore, the weld surface became rough.

stable in liquid bridge mode. However, due to high power density, the keyhole and molten pool stability were deteriorated which led to relatively drastic fluctuation near the keyhole, as shown in Fig. 8(d).

Fig. 9. Wire melting and transfer behaviors in droplet transfer mode during DBFW with tandem configuration.

143

Optics & Laser Technology 91 (2017) 138–148

G. Ma et al.

process. Similar to the result of tandem configuration, when the droplet size increased to a certain value, it contacted with the molten pool and flew into it. The size of the droplet was much smaller than that of tandem configuration. Until then, a transfer period was finished. Although the heated zone of dual laser beams on the droplet was not observed from the images, it was deduced that two small keyholes were formed on the droplet. Fig. 11 shows the typical droplet transfer process during the LWFW process. At the initial stage, the single laser beam started to melt the wire and formed a big but unstable keyhole in the molten pool. However during the droplet growing stage, there was no direct laser beam acting on the droplet. The size of droplet increased mainly by absorbing the plasma radiation energy and molten pool radiation energy. A fierce oscillation from the bottom of wire tip to the upside was observed during the process, as shown in Fig. 11(b). With increasing size, the big droplet would oscillate to the below of laser beam and seriously boil away with generating lots of spatters, as shown in Fig. 11(c). Meanwhile, most part of the liquid metal in droplet was thrown into the molten pool, which disturbed the keyhole and molten pool stability significantly. Compared the droplet transfer processes with different beam configurations, the results showed that the droplet transfer became stable in side-by-side configuration with the smallest droplet size and the least oscillation. It proved that the undesirable droplet mode could be improved by proper dual beam configuration. Because the melted wire was transferred to the molten pool in the form of droplet, so the transfer period was an important factor evaluating the transfer behavior. In this work, the transfer period was defined by the time interval between the moment that the droplet transferring into the molten pool completely. In each case, ten time intervals were extracted and the calculated average value was considered as the final transfer period. The results are shown in Fig. 12. The transfer periods of different beam configurations were 92 ms, 159 ms and 41 ms for single beam, tandem configuration and side-byside configuration, respectively. It indicated that side-by-side configuration could shorten the transfer period significantly, while the tandem dual beam configuration prolonged the period. This result

Based on the above analyse, the additional laser beam could influence the wire melting and transfer behaviors even the whole welding stability significantly. It showed positive or negative effect on the welding stability, which was controlled by the type of the dual beam configuration.

3.3. Effects of beam configurations in droplet mode Liquid bridge transfer mode was an optimal transfer mode with stable welding process. However, during practical welding process, this transfer mode was probably disturbed and changed into droplet transfer mode due to inaccurate wire feeding position. Droplet transfer mode was an important but unacceptable mode during laser welding with filler wire process, which provided an unstable welding process and bad weld appearance. In order to improve this case, dual beam laser welding was applied and effects of beam configurations on wire melting and transfer behaviors were also analyzed in detail. Fig. 9 shows the droplet transfer process during DBFW with tandem configuration. When the transfer process was relatively stable, one laser beam heated the molten pool and formed a keyhole; the other laser beam melted the wire tip and formed a small keyhole on the droplet, as shown in Fig. 9(a, b). With droplet growing up, the droplet started to oscillate up and down irregularly, as shown in Fig. 9(c, d). The red arrow in Fig. 9(d) showed the genaral oscillation direction. When the droplet increased to a certain size, it would contact with the molten pool and a “bridge” was estabilished during the violent oscillation process. Meanwhile, the liquid metal in the droplet transferred into the molten pool through the “bridge”, as shown in Fig. 9(e, f). After that, a new droplet would be formed and the transfer process was repeated periodically. The droplet transfer process during DBFW with side-by-side configuration is shown in Fig. 10. In this case, the dual laser beams heated and melted the wire tip directly at the both sides of filler wire, and then the melted wire retracted and grew up to a spherical droplet, as shown in Fig. 10(a, b). With the increase of absorbing energy, the droplet size increased. It should be noted that the position of droplet was always below the wire tip and few oscillation occurred in the

Fig. 10. Wire melting and transfer behaviors in droplet transfer mode during DBFW with side-by-side configuration.

144

Optics & Laser Technology 91 (2017) 138–148

G. Ma et al.

Fig. 11. Wire melting and transfer behaviors in droplet transfer mode during LWFW.

3.4. Analysis of welding stablity in droplet mode For droplet transfer mode, the droplet transfer behavior is determined by the forces applied on the droplet. The droplet is affected mainly by four forces during single/dual beam laser welding with filler wire: the gravity FG, the surface tension FS, the metal vapor/plasma ejecting force FV, and the evaporation recoil force FR. Among these forces, the gravity FG and the evaporation recoil force FR are detaching forces for droplet transfer, while the surface tension FS and the metal vapor/plasma ejecting force FV are restraining forces. The characteristic of each force is analyzed firstly. The gravity FG is defined as follows:

4 FG = πR3ρd g 3

Fig. 12. Average transfer periods in droplet transfer mode with different beam configurations.

(1)

where R is the droplet radius, ρd is the density of the droplet metal, g is the gravity accelation. So with increasing size of droplet, the gravity FG increases. The action direction of FG is downward to the surface of the workpiece. The surface tension FS is calculated as follows:

further confirmed the effect of side-by-side configuration on improving the welding stability in droplet transfer mode. The differences of transfer processes were also showed through the weld formation. Fig. 13 shows the weld appearances and cross sections in droplet transfer mode with different beam configurations. Obvious striations were observed in all weld appearance firstly. The striation width was also measured accurately by calculating the average of several striations. Their average widths were 1.61 mm, 2.82 mm and 0.79 mm for single beam, tandem configuration and side-by-side configuration, respectively. It indicated that the striation width was proportional to the transfer period and the shorter transfer period would lead to a better weld appearance. For the cross section, there was under fill defect at the bottom of weld for tandem configuration. This was attributed to that no enough energy was applied on the molten pool during the droplet transfer process. The other cross sections were relatively complete. Compared all these welds, the weld with side-byside configuration had the optimal weld formation, which significantly improved the weld formation of LWFW. It was directly proved that side-by-side configuration could improve the welding stability of droplet transfer mode.

FS =2πσw rw

(2)

where σw is the surface tension coefficient, rw is the radius of filler wire. According to the formula, the value of surface tension FS is considered to be a constant. The action direction of FS is upward along the wire axis. The metal vapor/plasma ejecting force FV is generated from the high speed metal vapor acting on the droplet which is ejected from the keyhole. This force can be considered as the drag force applied on a droplet immersed in the metal vapor/plasma. Similar to the action mechanism of plasma drag force in gas metal arc welding, FV is calculated as follows [20]:

⎛ ρ Vv 2 ⎞ FV =CD AP ⎜ v ⎟ ⎝ 2 ⎠

(3)

where CD is the drag coefficient of metal vapor, Ap is the projected area on the plan perpendicular to the flow direction of metal vapor, ρv is the density of metal vapor and Vv is the ejecting speed of the metal vapor. 145

Optics & Laser Technology 91 (2017) 138–148

G. Ma et al.

Fig. 13. Weld formation with different beam configurations in droplet mode: (a) tandem configuration, (b) side-by-side configuration, and (c) single beam.

single beam laser welding. Based on the above observation, when a laser beam was applied on the droplet, a small keyhole was also created, as shown in Fig. 9(b). So a reaction force applied on the droplet is also generated on the opposite direction of vapor/plasma ejection, which is the evaporation recoil force FR. The force is described as follows [21]:

Ap is given as follows:

Ap =πRp

2

(4)

where Rp is the radius of projected area if the droplet is spherical. It can be indicated that a higher power will produce more amount and higher speed metal vapor/plasma, which leads to a larger FV. Furthermore, due to the force is proportional to the projected area Ap, so the force is also related to the size of droplet. The action direction of FV varies with different beam configurations, which is related to the relative position between metal vapor/plasma and droplet, as shown in Fig. 14. The metal vapor/plasma between the two red dashed line arrow is the part applied on the droplet. The blue line arrow shows the action direction of FV. For tandem configuration, there is an angle between the direction of FV and the wire axis. For single beam laser welding, the action direction is nearly vertical to the top surface of workpiece. The metal vapor/plasma was usually unsteady, so the action directions even the values of this force are considered to be variable during the welding process, especially for

⎛ T − TLV ⎞ 2 FR=0. 54P0exp ⎜ΔHLV ⎟ πrk ⎝ RTTLV ⎠

(5)

where P0 is the atmospheric pressure, R is the universal gas constant, ΔHLV is the latent heat of evaporation, TLV is the boiling point of the base metal, T is the liquid temperature of droplet, and rk is radius of the acting area taken as the radius of keyhole in this work. The action direction of FR is also downward to the surface of the workpiece. The force analyse of droplet transfer behaviors in different beam configurations are schematically analyzed and compared. The results are shown in Fig. 15. Fig. 15(a) shows the droplet transfer process and the forces applied

Fig. 14. Action direction of FV with different beam configurations: (a) tandem configuration, (b) single beam.

146

Optics & Laser Technology 91 (2017) 138–148

G. Ma et al.

Fig. 15. Force analyse of droplet transfer behaviors with different beam configurations: (a) droplet transfer mode, (b) tandem configuration, (c) side-by-side configuration, and (d) sigle beam.

possibility that the droplet oscillated to the underside of laser beam followed by the droplet explosion and transfer into molten pool, when the transfer process was finished. By observing the droplet transfer process, it should be noted that the droplets were not transferred in the form of a complete droplet in the present work. This phenomenon indicated that the droplets did not reach the equilibrium condition of forces, which was due to that the radius of droplet was larger than the height h as shown in Fig. 3. It was calculated to be 0.87 mm when Wx was −1.5 mm. The critical droplet size in the equilibrium condition of forces was quantitatively analyzed. Before calculation, some necessary assumptions were made to simplify the process:

on the droplet in tandem configuration. It could be analyzed that there were four forces acted on the droplet and the force analysis as shown in Fig. 15(b). F was the resultant force and its magnitude and action direction would influence the droplet transfer behavior directly. Because the surface tension FSand evaporation recoil force FR were relatively steady in the process, so the transfer process was mainly determined by the gravity FG and the metal vapor/plasma ejecting force FV in this case. Firstly, with increasing droplet size, the gravity FG increased and the action direction of resultant force moved to the direction of the workpiece surface. This was beneficial to the droplet transfer. However, due to the instability of the metal vapor/plasma, the action direction of FV was probably changed, which affected the action direction of the resultant force and led to the droplet oscillation phenomenon as shown in Fig. 9(d). Until the droplet increasing to a certain size, the resultant force mostly pointed to the workpiece surface, and then it contacted wih the molten pool. After the liquid metal transferring to the molten pool completely, a transfer period was finished. For side-by-side configuration, the forces acted on the droplet differed with that of tandem configuration. In this situation, two small keyholes were generated on the droplet, so the evaporation recoil force FR was doubled compared to tandem confiuration, named FR1 and FR2, respectively. Meanwhile, the metal vapor/plasma ejecting force FV was neglected due to no obvious keyhole generated in the molten pool. The schematic of force analysis in side-by-side configuration is shown in Fig. 15(c). It could be seen that only the surface tension FSacted as restraining force. The direction of resultant was always toward to the workpiece surface, which explained the reason why droplet located below the wire tip. With increasing droplet size, it contacted to the molten pool and finished the transfer period. Therefore, with the help of double FR, the droplet could transfer into the molten pool with shorter time and smaller size than tandem confiruration. For single beam, the applied forces on the droplet were also changed due to the variation of beam configuration. In this case, there was no direct laser beam heating the droplet, so the evaporation recoil force FR was neglected. During the welding process, a bigger but more unstable keyhole was formed due to a higher power density than DBFW. Therefore, metal vapor/plasma ejecting force FV became larger and more unstable. The force even probably disappeared when the keyhole was completely closed during welding process. The force analysis of single beam is shown in Fig. 15(d). The magnitude and action direction of resultant force would be disturbed obviously, which led to the fierce oscillation during welding process. There was a high

(1) The shape of the droplets was spherical. (2) The height h was large enough that the droplet detached from wire tip completely. (3) The action direction of each force was fixed and the influence of oscillation was neglected. According to the static force balance model [20], when the detaching forces exceeded the restraining forces, the droplet detached from the wire tip. The critical droplet sizes in different beam configurations were calculated with different force analyse, as shown in Fig. 15. The parameters and values used in the calculation are shown in Table 3. The action directions of FV, named β, were set as 45° and 90°with respect to the horizontal surface of the workpiece for tandem

Table 3 The parameters and values used in calculation.

147

Nomenclature

Symbol

Value

Density of liquid deoplet Gravity accelation Surface tension coefficient Radius of filler wire Drag coefficient Density of metal vapor Ejecting speed of the metal vapor

ρd g σw rw CD ρv Vv

Atmospheric pressure Universal gas constant Latent heat of evaporation Boiling point Liquid temperature of droplet Radius of the acting area

P0 R ΔHLV TLV T rk

7200 kg/m3 9.8 m/s2 1.872 N/m 0.0006 m 0.44 0.06 kg/m3 150 m/s (LWFW) 100 m/s (DBFW) 101325 Pa 8.314 J/(mol·K) 411210 J/mol 3375 K 2350 K 0.00013 m

Optics & Laser Technology 91 (2017) 138–148

G. Ma et al.

prosecc stability problem shown in droplet transfer mode. (4) The differences of droplet transfer behaviors in different beam configurations were explained by force analyse. The action direction and stability of the resultant force deternmined the entire transfer process. The existence of unsteady metal vapor/plasma ejecting force FV resulted in the droplet oscillation phenomenon in tandem and sigle beam configurations.

configuration and single beam, respectively, which depend on the practical observation results as shown in Figs. 9 and 11, and the analysis in Fig. 14. Becauce the temperature of droplet T was not uniform from the small keyhole to the droplet surface, so an average temperature 2350 K was chosen to calculate the evaporation recoil force FR. For tandem configuration, the critical condition of droplet transfer in vertical direction is given as follows:

FSsinα w+FV sinβ =FG+FR

Acknowledgements

(6)

The research was finically supported by the National Natural Science Foundation of China (No.51175115). The authors also express our gratitude to Peng Jin for his assistance in welding experiments.

For side-by-side configuration, the critical detaching condition is given as follows:

FSsinα w=FG+FR1+FR2

(7)

For single beam, the critical detaching condition is shown as follows:

FSsinα w+FV =FG

References

(8)

[1] R.W. Davies, H.E. Oliver, M.T. Smith, G.J. Grant, Characterizing Al tailor-welded blanks for automotive applications, JOM 51 (11) (1999) 46–50. [2] N. Longfield, T. Lieshout, I.D. Wit, T.V.D. Veldt, W. Stam, Improving laser welding efficiency, Weld. J. 86 (5) (2007) 52–54. [3] V. Schultz, T. Seefeld, F. Vollertsen, Gap bridging ability in laser beam welding of thin aluminum sheets, Phys. Proc. 56 (2014) 545–553. [4] D.F.U. Dilthey, W. Scheller, Laser welding with filler wire, Opt. Quant. Electron. 27 (1995) 1181–1191. [5] Z. Sun, M. Kuo, Bridging the joint gap with wire feed laser welding, J. Mater. Process. Tech. 87 (1999) 213–222. [6] J. Xie, Dual beam laser welding, Weld. J. 81 (9) (2002) (223-s-230-s). [7] M.G. Deutsch, A. Punkari, D.C. Weckman, H.W. Kerr, Weldability of 1.6 mm thick aluminium alloy 5182 sheet by single and dual beam Nd: yag laser welding, Sci. Technol. Weld. Joi. 8 (4) (2003) 246–256. [8] A. Haboudou, P. Peyre, A.B. Vannes, G. Peix, Reduction of porosity content generated during Nd: yag laser welding of A356 and AA5083 aluminium alloys, Mat. Sci. Eng. A 363 (1–2) (2003) 40–52. [9] J.M. Drezet, M.S.F. Lima, J.D. Wagnière, M. Rappaz, W. Kurz, Crack-free aluminium alloy welds using a twin laser process, Proc. Int. Inst. Weld. Cinference (2008) 87–94. [10] B.J. Aalderink, B. Pathiraj, R.G.K.M. Aarts, Seam gap bridging of laser based processes for the welding of aluminium sheets for industrial applications, Int. J. Adv. Manuf. Tech. 48 (1) (2009) 143–154. [11] A.S. Salminen, V.P. Kujanpa¨a¨, Effect of wire feed position on laser welding with filler wire, J. Laser Appl 15 (2003), 2003, p. 2. [12] A. Salminen, The efects of filler wire feed on the efficiency of laser welding, Proc. SPIE 4831 (2003) 263–268. [13] A. Salminen, The filler wire - laser beam interaction during laser welding with low alloyed steel filler wire, Mechanika 37 (4) (2010) 67–74. [14] W.U.H. Syed, L. Li, Effects of wire feeding direction and location in multiple layer diode laser direct metal deposition, Appl. Surf. Sci. 248 (1–4) (2005) 518–524. [15] Y. Yu, W. Huang, G. Wang, J. Wang, X. Meng, C. Wang, F. Yan, X. Hu, S. Yu, Investigation of melting dynamics of filler wire during wire feed laser welding, J. Mec. Sci. Technol. 27 (4) (2013) 1097–1108. [16] W. Tao, Z. Yang, Y. Chen, L. Li, Z. Jiang, Y. Zhang, Double-sided fiber laser beam welding process of T-joints for aluminum aircraft fuselage panels: filler wire melting behavior, process stability, and their effects on porosity defects, Opt. Laser Technol. 52 (2013) 1–9. [17] K. Takahashi, S. Katayama, A. Matsunawa, Observation of filler wire melting dynamics during CO2 laser welding of aluminum alloys and evaluation of weldability, Q. J. Jpn. Weld. Soc. 20 (2) (2002) 220–227. [18] J.Zhou, H.L.Tsai, Investigation of fluid flow and heat transfer in 3-D dual-beam laser keyhole welding, in: Proceedings of the ASME 2010 International Manufacturing Science and Engineering Conferencepp. 247–254, 2010. [19] M.J. Torkamany, A.F.H. Kaplan, F.M. Ghaini, M. Vänskä, A. Salminen, K. Fahlström, J. Hedegärd, Wire deposition by a laser-induced boiling front, Opt. Laser Technol. 69 (2015) 104–112. [20] Y.S. Kim, T.W. Eager, Analysis of metal transfer in gas metal arc welding, Weld. J. 72 (6) (1993) (269-s-278-s). [21] J.Y. Lee, S.H. Ko, D.F. Farson, C.D. Yoo, Mechanism of keyhole formation and stability in stationary laser welding, J. Phys. D: Appl. Phys. 35 (13) (2002) 1570–1576.

By taking the formulas (1−5) into Eqs. (6–8), respecitively, the critical droplet radii were calculated to be 2.67 mm, 2.28 mm and 3.93 mm for tandem, side-by-side and single beam, respecitively. The droplet transferred with side-by-side configuration had the minium size, which decreased the droplet radius from 3.93 to 2.28 mm compared to the traditional LWFW process. It agreed well with the observed results as shown in Figs. (9–11). Generally speaking, a larger droplet needed more time to grow up and transfer. However, compared tandem and single beam configuration, the critical size with single beam configuration was larger, but the transfer period was shorter than that of with tandem configuration. This was mainly attributed to the fact that the fierce oscillation occurred in single beam configuration and supplied more chances for droplet explosion and transfer into molten pool within a certain period of time. 4. Conclusions Butt joints of 2 mm thick stainless steel with 0.5 mm gap were successfully welded by using dual beam laser welding with filler wire technique. The wire melting and transfer behaviors with different beam configurations were studied detailedly. The main conclusions of the study can be summarized as follows: (1) In liquid bridge transfer mode, the transfer process was disturbed and a “broken-reformed” liquid bridge was observed in tandem configuration. This was attributed to that the vapor/plasma ejecting force induced by the additional laser beam affected the stability of liquid bridge. (2) The best welding stability was obtained in side-by-side configuration with more stable transfer behavior and proper fluid pattern compared to the other two configurations. An improved weld formation was also presented. (3) In droplet transfer mode, the wire melting and transfer behaviors varied obviously for different beam configurations in the form of transfer stability, transfer period, and critical droplet size. The transfer stability of side-by-side configuration was better than the other two configurations with the minium transfer period and critical droplet size. An acceptable weld formation was also generated, which indicated the superiority on improving the

148