Numerical Analysis of Molten Pool Dynamic Behaviors during High-Speed Oscillation Laser Welding with Aluminum Alloy

Abstract

Increased welding speeds are often used to achieve high welding efficiency. However, welding defects, including pores and humps, can easily be formed under high-speed welding conditions. Therefore, a numerical calculation method is proposed to analyze the molten pool dynamic behaviors during high-speed oscillation laser welding with the aluminum alloy. The experiments on high-speed oscillation laser welding are conducted to confirm the simulated results, and both are found to be in good agreement. The distribution characteristics of the temperature field under the condition of a high welding speed are discussed. With the influence of the stirring process from the oscillation laser beam, the temperature gradient is reduced and fluid flow velocity is decreased. The fluid flow in non-oscillation laser welding is more intense than that in oscillation laser welding. It is found that the proposed method can be used to guide the selection of the optimal process parameters for improving welding efficiency and quality in the practical welding process.

1. Introduction

Aluminum alloy is attractive to various fields including rail transit, aerospace, automobile, and other related industries owing to its advantages, such as high strength and excellent corrosion resistance [1,2,3,4]. Laser welding is a precise, non-contact advanced technology and is highly flexible and accurate in joining aluminum alloy materials [5]. However, welding defects, including pores and hot cracks, can easily be formed in welds during aluminum alloy welding. Many advanced welding technologies are used to improve the temperature distribution and molten pool dynamic behaviors to suppress the formation of welding defects [5,6]. Welding efficiency is highly related to welding defect formation. Therefore, exploring the formation of welding defects during aluminum alloy laser welding at a high speed is practical for achieving high-quality welding.

Aluminum alloy laser welding has been studied widely by experimental methods to explore welding defect formation. Bunaziv et al. [7] analyzed the solidification, cracking, and porosity during laser beam and laser-arc hybrid welding of aluminum alloys. Ma et al. [8] obtained excellent welded joints without cracks in 7075 aluminum alloy laser welding. The cracking formation mechanism was revealed in their results. Indhu et al. [9] conducted a conduction welding process of a dual-phase steel and aluminum alloy by high-power diode laser. Their results showed that the mechanical properties of welded joints were greatly affected by the microstructure formed in the fusion region of the weld. Xu et al. [10] established a vacuum chamber with an oscillating electromagnetic molten pool support system to study the A5083 aluminum alloy laser welding process. Their observed results showed that the weld width in the upper zone and weld porosity were mainly affected by the ambient pressure. Kang et al. [11] analyzed the influence of beam intensity distribution on the weld during aluminum alloy butt welding. It was found that welding defects were reduced by the modulated beams, and a funnel-shaped weld was formed. Oana et al. [12] used filler material to adjust the chemical composition of the molten pool in the laser welding of some alloys with aluminum bases of the 6xxx series and discussed the effects of the additive materials on the macro- and microstructure in the welded joints. Herinandrianina et al. [13] found that the oscillating laser beam was an innovative technique that led to an important improvement in laser welding quality. Their results indicated that the oscillation amplitude had a great influence on the aluminum alloy welding quality. Jiang et al. [14] determined that the stronger stirring effects from the oscillating laser beam could capture the bubbles in the molten pool to greatly reduce the pores.

Numerical simulation is useful for revealing the underlying physics during the welding process. To understand the formation process of welding defects, the numerical calculation method has been introduced for aluminum alloy laser welding. Liu et al. [15] established a three-dimensional model considering the magnetohydrodynamics and keyhole-free surface to analyze the influence of magnetic field orientation on the suppression of weld porosity. They found that the long molten pool and small penetration depth were formed under the conditions of the magnetic field X-axis and magnetic field Y-axis, and the weld porosity defects were reduced greatly. Chen et al. [16] investigated the influence of oscillating frequency on weld formation and grain structure by numerical simulation. Their results showed that the frequency increase led to a decrease in actual heat input and, hence, the keyhole depth. Ke et al. [17] explored keyhole-induced porosity formation during welding with an oscillating laser beam by the numerical method, and the porosity formation mechanism was revealed. Beiranvand et al. [18] determined the relationship between the effective laser absorption coefficient of four Al alloys and, in parallel, electron-probe microanalyzer (EPMA) techniques by numerical simulation to obtain the Mg losses of the weld metals. Their analyzed results showed that Mg evaporation and the effective laser absorption coefficient were increased with the increase in laser pulse energy. Yin et al. [19] adopted numerical simulation modeling to study the laser mirror welding of aluminum alloys, and the characteristics and evolution mechanism of porosity were revealed. Their results showed that the large process porosities near the fusion line could be divided into two types.

However, the temperature field and fluid flow behaviors in high-speed aluminum alloy laser welding have rarely been studied in previous research, especially high-speed oscillation laser welding. Therefore, a numerical calculation method is proposed to analyze and discuss the temperature field distribution and molten pool dynamic behaviors in oscillation laser welding with aluminum alloy in high-speed welding. The cross sections of welds from non-oscillation laser welding and 8-shaped oscillation laser welding are compared. The dynamic behavior characteristics are extracted, based on the simulated results. It is found that the obtained simulation results are close to the experimental results. Therefore, the proposed method is validated and efficient for achieving an excellent weld with a high welding speed.

2. Mathematical Modeling

2.1. Governing Equations

Laser welding is a complex process in which the base material interacts with a laser beam to melt and achieve the joining of materials. It involves many complex physical phenomena including laser radiation, heat conduction, heat convection, etc. The schematic of the non-oscillation and 8-shaped oscillation laser welding processes is shown in Figure 1.

The heat and mass transfer phenomena involved in non-oscillation and 8-shaped oscillation laser welding can be expressed as follows [20,21]:

Continuity equation:

$$\nabla \cdot \overrightarrow{V}=0$$

(1)

Momentum equation:

$$\frac{\partial (\rho \overrightarrow{V})}{\partial t}+\nabla \cdot (\rho \overrightarrow{V}\overrightarrow{V})=-\nabla P+\nabla \cdot (\mu \nabla \overrightarrow{V})-\frac{\mu }{K}\overrightarrow{V}+\rho \overrightarrow{g}\beta (T-T_{ref})+S_M$$

(2)

Energy equation:

$$\frac{\partial (\rho H_{thc})}{\partial t}+\nabla \cdot (\rho \overrightarrow{V}{H}_{thc})=\nabla \cdot \left(k\nabla T\right)+S_E$$

(3)

where $\overrightarrow{V}$, $\rho$, t, and P are velocity, density, time, and pressure, respectively. μ is viscosity, and $\overrightarrow{g}$ is gravitational acceleration. $\beta$ is the thermal expansion coefficient. T, $T_{ref}$, and K are the workpiece temperature, reference temperature, and Carman–Kozeny coefficient. $H_{thc}$ is the enthalpy. k, $S_M$, and $S_E$ are thermal conductivity and source terms of momentum equation and energy equation.

The Carman–Kozeny coefficient, K, can be calculated as [22]:

$$K=K_0\frac{F_l^3+C_0}{{\left(1-F_l\right)}^2}$$

(4)

where $K_0$ is an empirical constant and $C_0$ is a small constant for convergence purpose. $F_l$ is the liquid fraction which can be described as [23]:

$$F_l=\{\begin{array}{cc}0\hfill & \hfill T<T_{sol},\\ \frac{T-T_{sol}}{T_{liq}-T_{sol}}\hfill & \hfill T_{sol}\le T\le T_{liq},\\ 1\hfill & \hfill T>T_{liq}.\end{array}$$

(5)

where $T_{sol}$ and $T_{liq}$ are the solidus temperature and liquidus temperature.

2.2. Laser Energy Absorption

During the laser welding process, the laser beam energy is transferred to the specimen. In the current research, the Gaussian rotating-body heating source model is consistent with the profile of the weld after multiple attempts. Hence the Gaussian rotating-body heating source model is adopted to describe the laser beam absorption. The heat flux of the Gaussian rotating-body heating source, $q\left(x,y,z\right)$, can be expressed as [24]:

$$q\left(x,y,z\right)=\frac{9\eta Q}{\pi R_0^{2}H\left(1-e^{-3}\right)}{e}^{\left(-\frac{9\left(x^2+y^2\right)}{R_0^2\ln (H/(h-z))}\right)}$$

(6)

where Q and $\eta$ are the laser power and laser absorption rate. x, y, and z indicate the coordinates of the Cartesian coordinate system in three directions. $R_0$ is the upper-surface radius of Gaussian rotating-body heating source. $H$ is the depth of the Gaussian rotating-body heating source and h is the thickness of the welding specimen.

2.3. The VOF Method

The volume fraction, F, is used to describe the change in phases in the computational domain. The volume fraction for the n-phase fluid is F_n, which is defined as follows. If F_n = 0, there is no n-phase fluid in the cell. If F_n = 1, the cell is filled with the n-phase fluid. If 0 < F_n < 1, the cell contains the n-phase fluid and other phase fluids. The volume fraction, F, can be calculated as [25]:

$$\frac{\partial F}{\partial t}+\nabla \cdot \left(\overrightarrow{V}F\right)=0$$

(7)

2.4. Boundary Conditions

Considering the energy loss caused by convection and radiation, the boundary condition for the surfaces of the base material is expressed as [26,27]:

$$k\frac{\partial T}{\partial \overrightarrow{n}}=-h_{con}\left(T-T_{amb}\right)-\epsilon \sigma \left(T^4-T_{amb}^4\right)$$

(8)

where $h_{con}$ is the convection coefficient and $\epsilon$ is the radiation emissivity. $\sigma$ is the Stefan–Boltzmann constant. $T_{amb}$ is the ambient temperature.

The Marangoni effect in the molten pool is taken into account, and the corresponding Marangoni force can be expressed as [28]:

$${\tau }_{mar}=\frac{\partial \gamma }{\partial T}\frac{\partial T}{\partial \overrightarrow{s}}$$

(9)

where ${\tau }_{mar}$ is the Marangoni force. $\gamma$ and $\overrightarrow{s}$ are the surface tension coefficient and unit area, respectively.

2.5. Numerical Implementation

The finite volume method is selected to solve the numerical models for the non-oscillation and 8-shaped oscillation laser welding processes by the computational fluid dynamics software, Fluent 19.2. The calculation time step is set according to the Courant number to ensure the stability of the solution process. The convergence of the iterative calculation process is judged as the calculating residuals. When the residual value is less than the standard convergence value, the iterative calculation process can be terminated. The corresponding thermophysical parameters of the 6061 aluminum alloy are determined by calculation software JMatPro and other published literatures and tabulated in

Table 1. The corresponding thermophysical parameters of 6061 aluminum alloy [29,30,31].

Name	Symbol	Value	Unit
Density	$\rho$	2660	kg/m3
Thermal expansion coefficient	$\beta$	2.8 × 10−5	1/K
Solidus conductivity	$k_{sol}$	235	W/(m·K)
Liquidus conductivity	$k_{liq}$	95	W/(m·K)
Solidus temperature	$T_{sol}$	873.15	K
Liquidus temperature	$T_{liq}$	915.15	K
Fusion latent heat	$h_l$	3.9 × 105	J/kg
Convection coefficient	$h_{con}$	20	W/(m2 K)

.

3. Experiments

3.1. Material

During the experimental processes of the non-oscillation and 8-shaped oscillation laser welding, 6061 aluminum alloy is selected as the base material. The corresponding chemical composition is shown in

Table 2. Chemical composition (wt.%) [32].

Si	Fe	Cu	Mn	Mg	Zn	Ti	Cr	Al
0.56	0.70	0.30	0.89	0.93	0.25	0.15	0.04	Bal.

. The dimensions of the specimens prepared for the welding process are 150 mm × 100 mm × 2 mm. Before welding, the alumina and oil stains on the welding specimen surfaces are removed. To avoid the aluminum alloy surface being oxidized again, the welding experiments are carried out immediately after removing the alumina and oil stains.

3.2. Laser Welding System

The experimental processes for the non-oscillation and 8-shaped oscillation laser welding are conducted. The oscillation laser is a fiber laser and from the galvanometer scanner. During the welding process, argon is adopted as the side-blow shielding gas, with a flow rate of 20 L/min to protect the welding region on the welding specimens. During the experimental process, the laser power and focal position are set as 1500 W and 0 mm. The effects of welding speeds on the weld morphologies in the 8-shaped oscillation laser welding are analyzed. The selected welding speeds are 3, 5, 7, 9, 11, 13, and 15 m/min in the 8-shaped oscillation laser welding. The oscillation frequency and amplitude are set as 120 Hz and 1.2 mm for the whole 8-shaped oscillation laser welding process. In the non-oscillation laser welding, the welding speed is set as 15 m/min.

3.3. Experimental Results Analysis

The weld morphologies formed under the conditions of different welding speeds are significantly different in the 8-shaped oscillation laser welding process. The weld morphologies in the cross sections and on the upper surfaces are extracted, and the effects of welding speeds on the weld morphologies are analyzed after the 8-shaped oscillation laser welding experiments.

The weld morphologies in the cross sections under the conditions of different welding speeds are shown in Figure 2. In Figure 2, it can be seen that the cross sections of the 8-shaped oscillation laser welding are wide in the upper part of the weld and narrow in the bottom part of the weld. The base material is fully penetrated under the conditions of welding speeds of 3 and 5 m/min, as shown in Figure 2a,b, while the base material is not fully penetrated under the conditions of welding speeds of 7, 9, 11, and 13 m/min, due to the reduction of the laser line energy, as shown in Figure 2c–f.

The weld morphologies on the upper surfaces under the conditions of different welding speeds are shown in Figure 3. It can be seen that the width of the weld region is almost uniform. Under the conditions of welding speeds of 3, 5, and 7 m/min, the upper surfaces of the formed welds are better quality, as shown in Figure 3a–c, while the obtained weld upper surfaces exhibit the “w” shape, and the uniformity of the width has been decreased under the conditions of high welding speeds of 9, 11, and 13 m/min, as shown in Figure 3d–f.

The influences of welding speeds on the weld morphologies are shown in Figure 4. The weld width decreases with the increase in welding speed from 3 m/min to 11 m/min, and the weld depth decreases with the increase in welding speed from 5 m/min to 13 m/min, as shown in Figure 4a. This is because the laser line energy decreases with the welding speed increase. From Figure 4b, the weld depth–width ratio decreases with the increase in welding speed from 5 m/min to 13 m/min, which indicates the weld quality is poor when the welding speed of 8-shaped oscillation laser welding of 6061 aluminum alloy is high.

4. Results and Discussion

4.1. Model Validation

The non-oscillation and 8-shaped oscillation laser welding experiments with aluminum alloy are carried out to validate the numerical calculation model. The process condition for the non-oscillation laser welding with the aluminum alloy is laser power of 1500 W, welding speed of 15 m/min, and focal position of 0 mm. The 8-shaped oscillation laser welding is conducted under the condition of laser power of 1500 W, welding speed of 15 m/min, focal position of 0 mm, oscillation frequency of 120 Hz, and oscillation amplitude of 1.2 mm. The experimental and calculated results for the welds in cross sections are shown in Figure 5. The calculated width and depth are kept consistent with the experimental results during the non-oscillation laser welding process, as shown in Figure 5a. The relative errors for the width and depth are 2.6% and 4.7%. Figure 5b shows the width and depth of welds from the 8-shaped oscillation laser welding with aluminum alloy, and both of them are found in good agreement. The relative errors for the width and depth are 6% and 14.6%, which are greater than those in the non-oscillation laser welding. This is because the laser absorption keeps changing induced by the oscillation laser beam during the welding process. Therefore, the proposed numerical calculation method is effective for the laser welding of aluminum alloy with a high welding speed.

4.2. Temperature Distribution of Molten Pool

Figure 6 indicates the temperature distribution on the top surface of the molten pool during the non-oscillation and 8-shaped oscillation laser welding with the aluminum alloy. In the molten pool of non-oscillation laser welding, the laser spot region with the high temperature is moving along the welding direction, as shown in Figure 6a. The obtained molten pool is elongated in the rear part of the laser spot region due to the fast-moving speed under the condition of high-speed welding. The length of the molten pool is shown as greater than its width. In Figure 6b, the practical movement direction of the laser is the combination of the welding direction and the oscillation direction of the laser. Compared to the molten pool from the non-oscillation laser welding, the formed molten pool of oscillation laser welding is found to be in different profile which is affected by the 8-shape oscillation path of laser beam. The molten pool width is increased under the oscillation amplitude effect, and hence the corresponding weld penetration is reduced under the certain absorbed laser energy, as shown in Figure 5b.

The temperature curves along the centerline under different oscillation amplitude conditions are extracted, as shown in Figure 7. It is found that the maximum temperature during the non-oscillation laser welding with the aluminum alloy (oscillation amplitude, 0 mm) is greater than that in the oscillation laser welding. The temperature value decreases from the position of the laser spot to the surrounding regions. During the non-oscillation laser welding with the aluminum alloy, the temperature peak is approximately 2974 K, which is above the evaporation temperature. Hence, the metal materials are evaporated during the welding process, and the formed weld is shown as the characteristic of high depth-to-width ratio.

4.3. Dynamic Behaviors of Molten Pool

The velocity distributions on the top surface of the molten pool during the non-oscillation and 8-shaped oscillation laser welding are shown in Figure 8. The flow velocity in the front part of the molten pool is greater than that in the rear part during the non-oscillation laser welding, as shown in Figure 8a. From the front to the rear part of the molten pool, the flow velocity shows the damping trend. Due to the influence of surface tension, the fluid flows from the center to the boundary. In Figure 8b, the molten pool is distributed along the oscillation laser direction during the 8-shaped oscillation laser welding process. The oscillation of the laser beam has a great stirring effect on the molten pool, and the fluid flow is in a complex mode, which can promote bubbles escaping and uniform temperature distribution.

To analyze the evolution behaviors in the 8-shaped oscillation laser welding, the fluid flow velocity evolutions of the molten pool in different positions along the oscillation path (Figure 9a) are extracted, as shown in Figure 9. The blue arrow in Figure 9a indicates the moving direction of oscillation laser beam. The laser beam is initially on the top boundary, and the fluid flows from the center to the boundary of the molten pool, which is kept the same as that in the non-oscillation laser welding, as shown in Figure 9b. As the laser moves towards the rear, as shown in Figure 9c, the laser beam starts to change oscillation direction. After the laser beam approaches the oscillation center, the molten pool is moving towards the bottom right corner, as shown in Figure 9d. As the laser moves to the right boundary, the molten pool length is increased, as shown in Figure 9e. The molten pool distribution direction is along the oscillation laser path. After the laser arrives at the bottom position, as shown in Figure 9f, the shape of the molten pool and fluid flow is symmetrical on the welding centerline with that in the initial stage, which means half of the period of oscillation laser beam is finished, and the next half oscillation period will begin. The laser beam moves forward according to the combined direction of the oscillation laser and high-speed welding direction. Therefore, a similar 8-shape on the upper surface of the weld is formed (Figure 3) after the molten pool solidifies.

Since the laser beam is kept moving between the left and right sides of the molten pool during the oscillation welding process, the fluid flow evolutions in the different longitudinal sections are illustrated in Figure 10. It can be seen that the fluid flow direction on the left section is opposite to that on the right section, which provides the condition for the stirring effect formation, as shown in Figure 10a,c. The fluid flow in the central section is from the front to the rear part in the upper region of the molten pool, as shown in Figure 10b. The vortex is found in the molten pool.

Figure 11 indicates the fluid flow velocity curve along the centerline with different oscillation amplitudes. The maximum fluid flow velocity is distributed in the central region of the molten pool in which the laser beam is located. The fluid flow shows fluctuant characteristics. In addition, the fluid flow in the non-oscillation laser welding (oscillation amplitude, 0 mm) process is much more intense than that in the 8-shaped oscillation laser welding. This is because the temperature gradient is reduced and fluid flow velocity is decreased under the influence of the stirring process from the oscillation laser beam. This phenomenon is efficient for the formation of excellent welds and high-quality welded joints [29].

5. Conclusions

A numerical calculation method is proposed to explore the temperature field and dynamic behaviors of the molten pool during high-speed laser welding with the aluminum alloy. The experimental and simulation results are found to be in good agreement. The temperature distribution characteristics of the molten pool with high-speed movement are analyzed. The dynamic behaviors of the molten pool from the non-oscillation and oscillation laser welding are compared based on the calculated results. It is found that the fluid flow in non-oscillation laser welding is more intense than that in 8-shaped oscillation laser welding. This is because the temperature gradient is reduced and fluid flow velocity is decreased under the influence of the stirring process from the oscillation laser beam. The 8-shaped oscillation laser welding makes it easier to achieve the uniform temperature distribution and stable flow field, which is beneficial for obtaining an excellent weld with a high welding speed.