Laser Keyhole Research Articles

Root formation and mechanical properties in laser keyhole welding of 15 mm thick HSLA steel

Deep penetration laser welding is promising in joining thick (> 10 mm) steel sections. Focused laser beam by drilling vapour cavity, the keyhole, generates deep and narrow welds. Full penetration single-pass joining has a persistent problem with root quality where humping is one of the most frequent imperfection. This strongly hampers the use of high-power laser for thick plate welding. A 16 kW disk laser was used for single-pass welding of 15 mm thick plates in a butt joint configuration. Root humping occurred within a wide range of welding parameters. This provides narrow processing window. By adding an arc source to the laser beam process, the tendency of root humping increases. To achieve humping-free welds and consistent root quality over length, a delicate balance of process parameters is required. High heat input (> 0.50 kJ/mm) was positive to achieve a combination of low hardness (< 325 HV) and good Charpy toughness at -50 °C (> 50 J).

Real-time monitoring of laser keyhole welding penetration state based on deep belief network

In laser keyhole welding processes, the dynamic behavior of the keyhole and molten pool is associated with the penetration state. In this paper, a deep learning-based monitoring system is employed to monitor the penetration state by inspecting the keyhole and molten pool. An image processing method based on parabola template and half molten pool is proposed to extract the boundary contour of the molten pool, which can effectively reduce the interference of metal vapor plume. 12 kinds of morphological features of the keyhole and molten pool are defined and extracted, especially 6 molten pool tail features. A data processing method is proposed to reduce the amount of data by 4/5 to avoid overfitting and remove abnormal data. The 12 variance features obtained through data processing can effectively increase the feature depth and improve the monitoring accuracy. A deep learning framework based on deep belief network (DBN) is established to model the relationship between the 24 features and their corresponding penetration states. The proposed framework achieves higher accuracy and stronger robustness in monitoring penetration states by comparing with the backpropagation neural network, support vector machine, decision tree, k-nearest neighbors and random forest. The results show that this proposed monitoring method can accurately and quickly judge the penetration state according to the keyhole and molten pool images obtained by the on-line monitoring system.

Effect of kerf characteristics on microstructures and properties of laser cutting–welding of AA2219 aluminum alloy

The purification of laser keyhole is an important means to remove the oxide layer of laser-MIG hybrid welding of aluminum alloy. Four common auxiliary gases (argon, nitrogen, air and oxygen) were used to prepare kerf with different characteristics, which were defined as argon-weld, nitrogen-weld, air-weld and oxygen-weld. The results show that the remelted layer has little effect on the weld morphology. Microstructures characteristics of Argon-welds and nitrogen-welds were the same as machined welds, while the narrow columnar grain zone and coarsen grains exist in air-welds and oxygen-welds. When the thickness of kerf remelted layer increases, the width of columnar grain zone is reducing and the diameter of equiaxed grain is increasing. The argon-weld is with the best tensile property. When the thickness of remelted layer (d) is less than 70 μm, the tensile strength and elongation of argon-weld can reach 278 MPa and 3.1% respectively, which equivalent to that of machined weld. The tensile property of nitrogen-weld is slightly lower than that of argon-weld because of aluminum nitride impurities. The tensile properties of air-weld and oxygen-weld are low because of a large number of oxide impurities and grain coarsening. Regardless of the types of kerf, when the d is less than 136 μm, the tensile strength reaches more than 250 MPa, which meets the application requirements.

Real-time laser keyhole welding penetration state monitoring based on adaptive fusion images using convolutional neural networks

Real-time laser keyhole welding penetration state monitoring based on adaptive fusion images using convolutional neural networks

A novel high-efficient welding technology with rotating arc assisted by laser and its application for cryogenic steels

To obtain high efficiency and good weld quality for medium thick steel plates, a novel welding technology of hybrid arc laser welding was proposed through rotating arc assisted by laser. Rotating arc provides heat to melt feeding wire and groove for high deposition and good fusion. Laser is helpful not only to arc rotating, but also to molten pool mixing. The proposed welding technology was used in the application of 9% Ni steel with medium thickness in the field of LNG storage tanks. The 21 mm thick steel plates were successfully achieved with only 2 welding passes. The weld formation was well-qualified, and no defects can be observed. The droplet transfer, keyhole and molten pool flow were investigated with high-speed camera and electric signal analysis. Experimental results showed that the droplet transferred in periodic alternation pattern of rotating spray transfer mode and short-circuit transfer mode. The laser keyhole kept stable during the droplet transfer process. The molten pool was fully mixed under the coupling effect of the arc, droplet and laser beam. The weld metal was in homogenous composition. Low temperature (−196 °C) impact energy of weld metal center and fusion line were much higher than related requirements. Obvious dimples on the fracture surface indicated that the fracture mechanism was in ductile fracture mode. Research results provided a theoretical basis for efficient construction and safe service of LNG containers made by cryogenic Ni steel.

Deep-learning-based real-time monitoring of full-penetration laser keyhole welding by using the synchronized coaxial observation method

In laser keyhole welding, the keyhole exhibits inherently unstable behavior, and the laser beam absorptance inside a keyhole varies rapidly. In this study, a real-time full-penetration laser keyhole welding monitoring system was established using a synchronized high-speed coaxial observation method in combination with deep learning models. Considering the images pertaining to the simultaneous observation of the top and bottom surfaces of the welding process, an object detection model (YOLOv4) was used to automatically measure the keyhole top and bottom apertures. The optimized model exhibited mean intersection over union accuracies of 98.23% (top) and 95.6% (bottom) and had a prediction speed of 156 fps. For 2-D images involving the measured keyhole top and bottom apertures, ResNet-34, which is a representative image classification AI model, was employed with an image regressor to predict the laser beam absorptance inside a keyhole; the model achieved an R 2 accuracy of 99.76% with a prediction time of 1.66 s for 740 keyhole geometries. The keyhole variations and absorptance fluctuated considerably during the welding progress, and the absorptance increased as the number of opened keyhole bottom apertures decreased, area of the keyhole bottom aperture reduced, and tilting angle increased. When a defect was generated, the laser absorptance declined rapidly as the keyhole bottom aperture size increased. Moreover, the width of the melt pool dramatically reduced.

Deep-learning approach for predicting laser-beam absorptance in full-penetration laser keyhole welding.

Laser-beam absorptance in a keyhole is generally calculated using either a ray-tracing method or electrodynamic simulation, both physics-based. As such, the entire computation must be repeated when the keyhole geometry changes. In this study, a data-based deep-learning model for predicting laser-beam absorptance in full-penetration laser keyhole welding is proposed. The model uses a set of keyhole top- and bottom-aperture as inputs. From these, an artificial intelligence (AI) model is trained to predict the laser-energy absorptance value. For the training dataset, various keyhole geometries (i.e., top- and bottom-aperture shapes) are hypothetically created, upon which the ray-tracing model is employed to compute the corresponding absorptance values. An image classification model, ResNet, is employed as a learning recognizer of features to predict absorptance. For image regression, several modifications are applied to the structure. Five model depths are tested, and the optimal AI architecture is used to predict the absorptance with an R2 accuracy of 99.76% within 1.66 s for 740 keyhole shapes. Using this model, several keyhole parameters affecting the keyhole absorptance are identified.

Material ejection attempts during laser keyhole welding

Abstract Material loss during keyhole mode laser welding often leads to surface imperfections that can negatively impact component performance. High-speed imaging identified four types of material ejection attempts: classical spatter, re-captured spatter, protuberances, and scalloping. The momentum attributed to the melt body, which is influenced by the keyhole properties, dictates whether the ejection attempt is successful or if the material is re-captured. The relationship between the dynamics of the keyhole and melt pool was elaborated in an extended systematic description of melt ejection attempts, which were then classified. Ejection attempts were often observed to be proceeded by a co-current swelling of the melt, adjacent to the keyhole, followed by a melt depression. The melt swell would elongate into a melt column with a concentration of momentum, where excessive momentum causes the melt to neck and separate into spatter. Trajectory determines if the spatter becomes a permanent fixture or re-incorporates into the melt body, with the latter having the possibility to cause further melt body disturbances leading to more ejection attempts. If the melt column fails to neck and separate, or an additional force acts upon the column, a protuberance or a scallop could then form. Keyhole and melt pool fluctuations were sometimes observed to be accommodated, avoiding material ejection. In these cases, a stable weld could be obtained with large variations in the dimensions of the melt pool and the keyhole.

Investigation on plume formation during fiber laser keyhole welding based on in-situ measurement of particles in plume

In this study, the particles in plume were in-situ measured based on the self-mixing effect of the semiconductor laser for revealing the plume formation during fiber laser keyhole welding. The plume consisted of two parts: the oscillating part at the bottom, and the slender part shaped like focused laser beam and with a decreasing ascending velocity. The diameters of particles in plume mainly distributed in the range of 0.3 μm∼0.65 μm (accounting for about 70 %). Along the laser beam direction, both the diameter and density of particles in plume decreased with the increase of the height. In the same horizontal measurement plane, the particle diameter distributed uniformly, and particle density decreased sharply when the monitoring plane further from the laser beam center. There are a great quantity of particles in the oscillating plume caused by eruption of the laser-induced vapor on the front keyhole wall, and that formation of the slender plume is related to the fact that when the oscillating plume is erupting along the laser beam direction, particles in it can enter the laser beam to move in the laser beam direction and to give out light after being heated by the laser beam.

Study of the effect of inter-pass temperature on weld overlap start-stop defects and mitigation by application of laser defocusing

Laser keyhole initiation and termination-related defects, such as cracking and keyhole cavities due to keyhole collapse, are a well-known issue in laser keyhole welding of thick section steels. In longitudinal welding, run-on and run-off plates are used to avoid this problem. However, such an approach is not applicable in circumferential welding where start/stop defects remain within the workpiece. These issues can hinder industry from applying laser keyhole welding for circumferential welding applications. In this paper, the effect of inter-pass temperature on laser keyhole initiation and termination at the weld overlap start-stop region was investigated. This study has identified that defects occurring within this region were due to laser termination rather than laser initiation because of keyhole instabilities regardless of the thermal cycle. The laser termination defects were mitigated by applying a laser defocusing termination regime to reduce the keyhole depth gradually and control the closure of the keyhole.

Enhancing coupling efficiency in laser keyhole welding of copper using femtosecond laser surface modification

To improve laser beam coupling to the copper surface for the formation of a keyhole, a method of changing the surface reflectance of copper using femtosecond laser patterning is proposed in this study. For 2 kW fiber laser welding of copper, a 2 mm × 5 mm pattern was fabricated on the copper surface at the welding start point, taking a maximum of 27 s to produce a suitable pattern. The surface reflectance of copper was subsequently reduced from 97% to as low as 3% (for a 1 μm wavelength) by femtosecond laser patterning based on laser-induced periodic surface structures and their variants. Moreover, when a low reflectance pattern was employed, it was possible to increase welding speed by up to 300%, and once the melt pool was fully established, its size remained relatively stable for the length of the weld line. However, owing to melt pool fluctuations and the large conductive heat loss of copper, the depth and width of the fusion zone decreased and the laser coupling to the copper surface was lost in several cases. This problem was overcome by placing the patterns at regular intervals along the welding direction, indicating that this method has long weld-length applications.

Laser Keyhole Brazing

Laser Keyhole Brazing

A study on laser keyhole welding in vacuum by computational fluid dynamics simulations with plume effect models

A vacuum environment results in a deep penetration depth in laser keyhole welding. In this study, the numerical simulations of laser keyhole welding in various vacuum conditions were performed with consideration for multiphysics phenomena. To consider the vacuum environment in numerical simulations, a modified recoil pressure model with the change in pressure and vaporization temperature according to ambient pressure and a laser power attenuation model by scattering and absorption due to the nanoparticles in the laser-induced plume was considered. For the scattering and absorption calculations, artificial coefficients obtained for a 1 atm condition were used, and then the coefficients in the other pressure conditions were calculated under assumptions suggested in this study. The simulation results using the models based on the assumptions could depict a deeper penetration in vacuum conditions considered in this study, and they were in good agreement with the experimental results in terms of weld cross section.

Determination of heat source model for simulating full penetration laser welding of 316 LN stainless steel by computational fluid dynamics

The selection of appropriate heat source model is vital in predicting the temperature distribution and visualizing the dynamic fluid flow behaviour during Laser keyhole welding of Type 316 LN stainless steel. Three-dimensional (3D) numerical models were designed utilizing ANSYS Fluent by employing different heat source models. Different heat source models considered include volumetric heat sources namely Goldak double ellipsoidal, 3D conical, a coalition of 3D conical and cylindrical heat source, 3D volumetric Gaussian and Rotary Gaussian. The simulation was carried out by solving the fundamental conservation equations of mass, momentum and energy. The predicted weld bead profile by different heat sources were compared with the actual weld bead obtained from the laser welding experiment on 0.0055 m thick 316 LN stainless steel (SS) plates. It was found that simulation results of the conical-cylindrical combined (CCC) model exhibited good agreement with actual weld bead dimensions and thermal data obtained during the experiment. Therefore, the conical-cylindrical heat source model is recommended for CFD simulation of full penetration laser welding of 316 LN stainless steel for studying the effect of variation of laser power and surface active element on the weld pool development.

Correlation between the spatters and evaporation vapor on the front keyhole wall during fiber laser keyhole welding

Severe spattering is one of the common problems during fiber laser keyhole welding. In this study, the correlation between the spatter formation and laser-induced evaporation vapor on the front keyhole wall (FKW) was studied by changing the welding speed during fiber laser keyhole welding. The spatters were mainly generated in front of and behind the keyhole mouth, and the diameters of spatters were mainly between 50~100 μm at low welding speed, while the diameters of spatter particles were mainly between 100~150 μm at high welding speed. As the welding speed was increased, the tilt angle of the FKW decreased gradually, and the main spatter generation position shifted from in front of the keyhole mouth to behind the keyhole mouth. In low welding speed, the spatter velocity was higher than that in high welding speed. Further analysis shows that, the recoil pressure of laser-induced evaporation squeezed the molten metal layer in the FKW to generate spatters in front of the keyhole mouth in low welding speed. In high welding speed, the decrease of the molten metal layer in the FKW led to the decrease of spatters in front of the keyhole mouth; however, the laser-induced evaporation vapor from the FKW was more liable to impact the rear keyhole wall to generate spatters behind the keyhole mouth.

Evolution behavior of laser welding in hybrid structure between open-cell aluminum foam and solid aluminum shell

Hybrid structure between porous core and solid shell can combine lightweight open-cell structure of porous materials with good strength of solid shell. Laser welding is considered as an effective process to join porous metal with solid metallic shell. In this study, the evolution behavior of laser welding by a continuous wave (CW) disk laser welding between open-cell aluminum alloy (cast AC2A alloy) and aluminum sheet (wrought AA6063 alloy) was investigated. X-ray transmission system and a high-speed camera were performed during laser welding to investigate the dynamic keyhole, porosity formation, plasma plume, and spatter formation. Laser keyhole became unstable during entering and passing the vacancies of porous aluminum, which consequently resulted in bubble formation in the molten pool.

Influence of the beam oscillation pattern and oscillation frequency on the temperature field in laser brazing with keyhole formation

Laser keyhole brazing is an opportunity to increase the process efficiency in laser brazing processes. Using small spot sizes increases the intensity and leads to the formation of a vapour capillary (keyhole) in the brazing material when a material-specific threshold value is exceeded. Due to multiple reflections/absorptions of the laser beam in the keyhole, the process efficiency increases in comparison with conventional brazing processes with single Fresnel absorption on the surface, especially when using high-reflectivity braze materials, such as aluminium-based or copper-based alloys. The energy must be distributed adequately by applying beam oscillation transversal to the brazing direction. In laser brazing processes, the temperature field in the interface between brazing and substrate material is a major factor. To analyse the effect of beam oscillation, it is assumed in this study that the temperature distribution at the surface of the melt pool is a suitable approximation for the temperature distribution at the interface to the substrate. Two key parameters are defined to quantify the temperature field referring to the homogeneity: the temporal local temperature-time curve and the temperature distribution transverse to the brazing direction. While the oscillation frequency influences the first mentioned parameter by decreasing the time interval between the local laser passes, the oscillation pattern affects the second parameter by adjusting the local actual beam velocity and its consistency.

Metal evaporation flux across Knudsen layer in laser keyhole welding of Al–Mg alloys with pressure balance condition method

This study aims to calculate the metal evaporation flux across the Knudsen layer inside the keyhole in laser welding of Al–Mg alloys. The pressure balance at the keyhole wall was analysed and used as the boundary condition in the calculation. Based on the extracted keyhole shape parameters from high-speed images, the surface temperature and gas pressure inside the keyhole were calculated, and the evaporation fluxes for both Al and Mg were acquired. The results show that the front wall of the keyhole absorbs more laser energy than the rear wall. The gas pressure inside the keyhole increases and the evaporation flux decreases with increasing keyhole depth. In addition, the evaporation flux of Mg is much higher than that of the base element Al, which can result in the loss of Mg inside the melt pool. The distribution of the Mg evaporation flux inside the keyhole agrees with the measured Mg loss in the welded section of the Al–Mg alloy. Thus, the calculation method of the metal evaporation flux inside the keyhole can be applied in future simulations of element loss in laser keyhole welding.

Study on effect of laser keyhole weld termination regimes and material composition on weld overlap start-stop defects

Abstract The effects of two different laser termination regimes; laser power ramp-down and laser defocusing, on the weld overlap start-stop region, have been studied on three different compositions of S355J2 + N grade steel plates. The plates have varying amounts of deoxidisers, such as manganese and silicon, to allow for the effect of dissolved oxygen on melt pool dynamics and defect formation to be studied. Observations show that gas entrapment, manifesting as porosity, is the main issue in both the laser termination regimes studied, due to melt pool instabilities related to keyhole termination. However, it is more severe during laser defocusing due to the formation of porosity clusters below a specific power density and melt pool size. The percentage of dissolved oxygen and deoxidisers was found to be vital and could be correlated to the response of the melt pool during laser processing. Steel, with a higher percentage of deoxidisers, such as manganese and silicon, would combine with the dissolved oxygen in the weld pool and thereby affect the surface tension and melt pool dynamics. The effect of lower dissolved oxygen would lead to a reduction in penetration depth, but more importantly, a reduction in porosity and porosity cluster formation, as observed for both laser termination regimes. Successful laser termination with defect-free weld overlaps at the start-stop region could be achieved using both termination regimes for the steel with a higher proportion of deoxidising elements.

Investigation of metal mixing in laser keyhole welding of dissimilar metals

Laser keyhole welding of dissimilar metals has broad applications in various industrial fields. However, dissimilar metal mixing in the molten pool often causes the formation of detrimental intermetallic compounds that can undermine the performances of the dissimilar metal joints. In this study, the metal mixing process in laser keyhole welding of dissimilar metals is investigated with a combination of experimental and modeling approaches. The parametric experimental study is conducted to reveal the effects of laser power, welding speed, and heat input on the metal mixing in the fusion zone. Ex-situ energy-dispersive X-ray spectroscopy element mapping is used to characterize the metal mixing status in the fusion zone. To investigate the underlying physics of the welding process, a numerical model is developed to simulate the heat transfer, fluid flow, and metal mixing in the molten pool. It is found that the recoil pressure contributes to an upward flow that pushes copper to migrate from the bottom of the molten pool to the top. Meanwhile, the Marangoni force generates one backward flow and two side vortices that facilitate the metal mixing. An increase of laser power increases both the recoil pressure and the Marangoni stress, which drives more copper to migrate upward and mix with aluminum. An increase of welding speed reduces the molten pool lifetime, which reduces the metal mixing by the fluid flow. The investigation of dissimilar metal redistribution provides insights regarding the formation of intermetallic compounds in the joints, which is valuable for the design and optimization of the welding process in different industries.