Molten Pool Width Research Articles

CNN-BiLSTM enabled prediction on molten pool width for thin-walled part fabrication using Laser Directed Energy Deposition

Laser Directed Energy Deposition (LDED) is a promising metal Additive Manufacturing (AM) technology capable of fabricating thin-walled parts to support some high-value applications. Accurate and efficient prediction on the molten pool width is critical to support in-situ control of LDED for part quality assurance. Nevertheless, owing to the intricate physical mechanisms of the process, it is challenging to designing an effective approach to accomplish the prediction target. To tackle the issue, in this research, a new data model-driven predictive approach, which is enabled by a hybrid machine learning model namely CNN-BiLSTM, is presented. High prediction accuracy and efficiency are achievable through innovative measures in the research, that is, (i) the CNN-BiLSTM model is designed and configured by addressing the characteristics of the LDED process; (ii) process parameters related to the deposition and heat accumulation phenomena during the LDED process are extensively considered to strengthen the prediction accuracy. Experiments for thin-walled part fabrication were conducted to validate and benchmark the approach. In average, 4.286% of the mean absolute percentage error (MAPE) was acquired, and the prediction time took by the approach was only 0.04% of that by a finite element analysis (FEA) approach. Compared to the LSTM model, the BiLSTM model and the CNN-LSTM model, MAPEs of the CNN-BiLSTM model were improved by 27.0%, 17.3% and 12.6%, respectively. It demonstrates that the approach is competent in producing good-quality thin-walled parts using the LDED process.

Towards hybrid laser-arc based directed energy deposition: Understanding bead formation through mathematical modeling for additive manufacturing

This study highlights three phenomena in directed energy deposition using hybrid laser-arc: bead formation, change in the bead shape because of laser-arc interaction, and formation of defects. Unlike the arc-based deposition, the shape of the hybrid laser-arc bead differs from the pre-conceived parabolic or cosine form, which necessitates a new bead model. The laser-induced flow and preheating of wire electrode widen the molten pool width and cause a depression of the top surface of the deposited bead. Consequently, the fixed critical offset distance for multi-pass deposition (e.g., 0.66 or 0.73 times of the bead width, based on the flat or tangent overlapping model for arc-based deposition) does not yield the optimal flat surface for the hybrid laser-arc deposition. In this work, first, a parametric study assesses bead formation and defect phenomenon for high-strength steel as a candidate material, followed by an evaluation of the existing bead models for the appropriateness of the shape of defect-free hybrid laser-arc beads. Results indicate that a parabola composited with a unimodal beta distribution - that accounts for the laser influence (power and laser-arc distance) on the bead shape - is more responsive to the bead shape variation and is substantially more accurate than the existing models. Increasing laser power and laser-arc distance significantly changes the bead shape and reduces the critical offset distance. The traverse speed is the most critical factor followed by welding current leading to defect-free bead deposition. It is envisioned that the fundamental understanding developed as a result of the present investigation will provide a direction for hybrid laser-arc based additive manufacturing of several material classes, many of which are not weldable and therefore not additively manufactured by arc-based deposition.

Coupling Analysis on Microstructure and Residual Stress in Selective Laser Melting (SLM) with Varying Key Process Parameters.

With the application of Selective Laser Melting (SLM) technology becoming more and more widespread, it is important to note the process parameters that have a very important effect on the forming quality. Key process parameters such as laser power (P), scan speed (s), and scanning strategy (μ) were investigated by determining the correlation between the microstructure and residual stress in this paper. A total of 10 group 316L specimens were fabricated using SLM for comprehensive analysis. The results show that the key process parameters directly affect the morphology and size of the molten pool in the SLM deposition, and the big molten pool width has a direct effect on the larger grain size and crystal orientation distribution. In addition, the larger grain size and misorientation angle also affect the size of the residual stress. Therefore, better additive manufacturing grain crystallization can be obtained by reasonably adjusting the process parameter combinations. The transfer energy density can synthesize the influence of four key process parameters (P, v, the hatching distance (δ), and the layer thickness (h)). In this study, it is proposed that the accepted energy density will reflect the influence of five key process parameters, including the scanning trajectory (μ), which can reflect the comprehensive effect of process parameters more accurately.

Effect of molten pool and vapor/plasma plume dynamic behavior of laser-induction hybrid welding on weld formation of S690QL steel

For overcoming the shortcomings of traditional welding methods, laser-induction hybrid welding (LIHW) was investigated. In this paper, high strength low alloy steel S690QL was selected to explore the improvement mechanism of LIHW on weld formation, mainly through high-speed photography, optical microscope and electron back-scattered diffraction (EBSD) techniques to compare the vapor/plasma plume and keyhole dynamic behaviors of laser beam welding (LBW) and LIHW. Both the characteristics of weld joint geometry and microstructure processed by LIHW technology were better than LBW. With increasing inductor output power, the changing period of vapor/plasma plume was accelerated, the spatter defects were inhibited, the molten pool length and width increased. At the same time, due to the better welding heat input absorption and utilization, most the LIHW weld joints contain more ferrite, lower bainite and martensite than LBW weld joint. Furthermore, the alternating magnetic field and the two eruption paths of upper keyhole and lower penetrated hole in LIHW processing could lead to the penetration of LIHW joints, while the LBW does not. Thus, the weld formation mechanism was revealed and compared with LBW, LIHW have better molten pool and vapor/plasma plume dynamic behaviors, which result in the improvement of weld formation.

Numerical simulation and microstructure formation mechanism of Ni-based coating fabricated by laser on copper plate

The coating on copper plate is difficult to be fabricated by laser. In order to obtain the optimal parameters of laser cladding coating, the temperature fields at different laser power and scanning speed were investigated and analyzed. The optimal parameters were selected with the depth and width of molten pool and compared with the experimental results. Then, the cooling rate, temperature gradient and solidification rate were calculated and analyzed with the coating’s microstructure. The results show that the optimal process parameters are 1500 W and 2 mm/s. And at these parameters, the temperature field is consistent with the morphology of cladding coating. From bottom to top of the coating, the corresponding cooling rates are 1056.7, 936.5 and 762.6 ℃/s, and the microstructure are elongated disordered dendrites, columnar dendrites and planar crystals, respectively.

Dynamic molten pool behavior of pulsed gas tungsten arc welding with filler wire in horizontal position and its characterization based on arc voltage

To explore the dynamic behavior of molten pool and to find the electric characteristic signals reflecting the weld penetration states and degree for the fixed-point wire-filled pulsed gas tungsten arc welding (GTAW-P) of thin pipes in 2G position, molten pool images from different directions and arc voltage were sensed. During a peak current period, critical penetration occurred, with the molten pool behavior that the topside molten pool height collapsed sharply and was approximately 0.5 mm, and then rose to approximately 1.5 mm during the next base current period. With the increase in backside molten pool width, the oscillation amplitude and frequency of topside molten pool during the base current period increased and decreased, respectively. The oscillation amplitude reached the maximum when over penetration occurred. Such molten pool behaviors could be extracted from the arc voltage signals. The change in average peak voltage (ΔU⁎) could reflect the occurrence of critical penetration. The average base voltage (Ub¯) and the fluctuation of base voltage (ΔUb) increased linearly with the increase in backside molten pool width, which are reliable characteristic signals for the weld penetration degree. The effects of wire filling and base current on the dynamic molten pool behavior and characteristic signals were investigated. Experimental results showed when the base current is 10 A ̶15 A, Ub¯ and ΔUb can effectively characterize the weld penetration degree.

Mesoscopic evolution of molten pool during selective laser melting of superalloy Inconel 738 at elevating preheating temperature

Due to the poor weldability of Inconel 738 alloy, defects (cracks, pores) always exist in the selective laser melting(SLM) process. In this paper, the molten pool evolutionof the single track, including the morphology, porosityand crack formation, wassystematically investigated by numerical simulation, andthe effects of different preheating temperature and process parameters were verified by experimental analysis. The detailed temperature-dependent thermal physical properties were applied into the Volume of Fluid (VOF) methodto simulate a powder bed computational domain which was established by Discrete Element Model (DEM).The simulation results show that the change of molten pool width by different laser power and preheating was not as significant as that of depth. The results also demonstrate that higher energy density and preheating temperature induced the generation of pores at the bottom of molten pool, which can be explained by the variation in cooling rates . Moreover, the change of temperature gradients in three-dimensional directions with the increase of preheating temperature was also investigated, which explained the reason for crack forming and extending through the center line. Finally, the evolution of molten pool at a higher preheating temperature1000℃ was simulated to further underpin the positive effect of preheating on preventing cracks formation.

Power density effect on the laser beam-induced eruption of spatters in fiber laser keyhole welding

Violent eruption of spatters is one of the major problems restricting fiber laser keyhole welding technology development. In this study, the fiber laser power density variation was attained by altering the laser power and laser spot diameter for analyzing its impact on the eruption of spatters. The severity of spatters’ eruption can be indicated by the number of spatters and the mass loss of the weld seam. The correlation between the number of spatters and laser power density can be positive (if the latter is adjusted by changing the laser power) or negative (if the laser power density is varied by changing the laser spot diameter). When the latter diameter is changed, it has a strong impact on the vapor induced by the laser on the front keyhole wall (FKW), the FKW inclination angle, and the molten pool surface tension. The number of spatters also positively correlates with the laser-induced vapor on the FKW or the molten pool width and negatively correlates with the FKW inclination angle. Reducing the laser spot diameter can increase the laser power density, enhancing the welding process stability and suppressing the eruption of spatters to a certain extent.

Effect of beam oscillation on intermetallic compounds and mechanical properties of steel/aluminum laser welded joint

Brittle Fe-Al compounds limit the performance of the steel/aluminum joint. Beam oscillation laser technology was performed to control the thickness, morphology and distribution of intermetallic compounds (IMCs) at interface. The results showed that beam oscillation significantly reduced the thickness of IMC layer. When oscillation frequency was 10 Hz, the needle-like phase embedded in aluminum matrix was obviously passivated in the circular oscillation mode, while the needle-like phase at interface tend to distribute parallel to the layer-like IMC, and formed a “bite” connection effect with aluminum matrix. In the linear oscillation mode, the needle-like phase disappeared, and smooth and flat interface was obtained. Compared with the non-oscillation mode, in the case of circular or linear oscillation mode, the width of molten pool was significantly increased, the crystal grains were refined, and the joint property was increased by 26% and 17% respectively, which were caused by the combined effect of mechanical bonding and metallurgical regulation. The influence mechanism of beam oscillation on the thickness, morphology and distribution of IMC at interface was revealed by high-speed photography and finite element simulation, which provides a new idea for the welding of steel/aluminum dissimilar materials.

Coordinated monitoring and control method of deposited layer width and reinforcement in WAAM process

Wire arc additive manufacturing (WAAM) has been used extensively in metal manufacturing and other fields because of its low cost and high efficiency. In the manufacturing process, WAAM technology is often unable to be further applied due to issues such as component dimensional accuracy, layered morphology, and metallurgical defects. To overcome these problems, online monitoring and process control of the welding are necessary. Based on the monitoring of weld width and reinforcement, a regression network for extracting the global information of molten pool is proposed, and an active disturbance rejection control (ADRC) is designed to adjust the welding current. To the best of our knowledge, this is the first time to realize the coordinated monitoring and control of the width and reinforcement of the deposited layer in the WAAM process. Experiments show that the method can obtain satisfactory molten pool width and reinforcement control accuracy, which provides a feasible way for monitoring and control the weld shape in the WAAM welding process.

Effects of different active fluxes on the molten pool behaviors of laser-welded thick stainless steel plates

Dynamic behaviors of molten pools during active laser welding of thick 304 stainless steel plates are monitored in real time by utilizing a high-speed camera system. The effects of various active fluxes on the dynamic variations of molten pools are analyzed, and the results are compared with the uncoated specimens. The research reveals that all the molten pools present periodic variations. Among them, the specimen coated with NaF active flux exhibits the shortest cycle of molten pool change, with a single cycle duration of approximately 0.5 s. The specimen coated with TiO2 active flux exhibits the longest cycle of molten pool change, and the single cycle duration is around 1.1 s. Moreover, the molten pools also show varying widths. The maximum molten pool width is found with the specimen coated with CeO2 active flux, whose value is 3.277 mm. Meanwhile, the minimum molten pool width is found with the specimen coated with NaF active flux, whose value is 1.799 mm. In addition, the coating of active fluxes also produces certain effects on molten pool stability. The specimen coated with NaF active flux presents intense emission of metal vapor during welding, drastic oscillation of the molten pool, and severe spattering of metal particles. Contrastively, the molten pool of the specimen coated with Cr2O3 active flux is rather stable. In conclusion, by selecting an appropriate active flux, the cycle of molten pool variation can be prolonged, the pool stability can be enhanced, and the occurrence of spattering and other defects can be suppressed.

Investigation of temperature distribution and melt pool microstructure in laser fusion welding of Inconel 625 superalloy

Nickel-based super alloy laser welding is of particular importance because of its numerous usages in the energy and aerospace industries. Measuring the temperature field is the basic criteria for conducting a qualitative evaluation of the weld joint. In this research, laser welding was experimentally investigated and the temperature field was measured. Measuring the temperature around the molten pool by varying the laser parameters such as nozzle distance, welding speed, laser power, and beam offset indicated a different heat field, resulting in changes in the molten pool's width and depth. Because of the high temperature of melting and low thermal conductivity coefficient of the Inconel 625 alloy, the measured temperature was large. Compared with the other parameters, the effect of enhancing the laser power on temperature increase around the molten pool was significant. The findings showed that, by increasing the laser power from 300 to 400 W, the temperature increased from 320 to 340 °C. Also, by increasing the nozzle distance from the surface of the workpiece to 2 mm, the temperature decreased from 300 to 200 °C.

Effect of welding speed, pulse frequency, and pulse width on the weld shape and temperature distribution in dissimilar laser welding of stainless steel 308 and brass alloy

Laser welding of brass and stainless steel alloys is of special importance due to its wide application in the industries related to energy production. In this study, laser welding of heterogeneous metals was performed in a laboratory. Measurements of the temperature around the molten pool showed changes in the welding conditions. By changing different parameters such as welding speed, frequency, and pulse width, different thermal gradients were obtained. The results showed that the formation of the molten pool was asymmetric, and it was mostly done by melting the brass alloy. Due to the lower melting temperature and the high heat transfer rate of the brass alloy, the measured temperature and the molten volume of this alloy were higher. The microstructure of the molten pool also included intermetallic compounds. Increasing the dimensions of the molten pool (width and depth) by raising the peak power and reducing the welding speed was more effective than other parameters. The microhardness results also indicated the higher weld strength of the brass alloy with stainless steel rather than pure copper. At the range of the investigated process parameters, the adjacent temperature near the molten pool was about 20% (30 °C) higher for the brass alloy, in comparison to stainless steel, when the pulse frequency and pulse width were changed.

Selective laser melting of Cu-inconel 718 powder mixtures

• Cu mixed with Inconel 718 power hinders the densification process during SLM. • Compared to pure Inconel 718, on increasing the Cu content in the feedstock powder it is observed. • A growing reduction of the molten pool width. • An intensifying frequency of lack-of-fusion defects resulting in a larger porosity. • A progressively larger loss of hardness. The selective laser melting of mechanically mixed Inconel 718 and copper powders is studied by processing feedstocks of different Cu content at various energy inputs. It was observed that with the increase of the Cu content the width of the molten pools decreases while the porosity of the samples increases. Lower values of energy input enhance the effects brought about by the addition of Cu. Decreasing the Cu weight fraction or, inversely, increasing the energy input, the tensile and yield strengths progressively increase due both to the reduced content of the softer component, such as the Cu powder, and to the decrease in porosity. Because of the porosity increase, the decrease in hardness and increase in thermal conductivity, registered on incrementing the copper content in the feedstock, are, respectively, greater and lower than expected. The obtained results show the possibility of manufacturing dense part by SLM using Cu-Inconel 718 mixed powders.

Numerical analyses of molten pool evolution in laser polishing Ti6Al4V

Abstract As one advanced manufacturing process, laser polishing has been widely used in precision machining of metal or nonmetal materials. However, the evolution mechanism of molten pool formed during polishing process has not been thoroughly studied. This paper studies the polishing process of Ti6Al4V with a continuous wave (CW) fiber laser under the top-hat distributed heat source. A two-dimensional (2D) axisymmetric numerical transient model is established to simulate the formation process of the surface of Ti6Al4V after laser heating to form a molten pool and then cooling. It reveals that the evolution mechanism of physical processes involved in heat transfer, heat radiation, heat convection, melting and solidification during the polishing process. Particularly, the capillary and thermocapillary regimes (Marangoni flow) play a key role in smoothing the free-form surface of molten pool. The freedom surface can be smoothed effectively before the solidification of molten pool. In order to verify the correctness of the simulation model and explore the influence of laser irradiation duration on the material, the surface topography of the polished sample was investigated. It is found that the errors of the depth and width of the remelting zone are respectively less than 8 % and 22 % by the comparison of simulation and experimental results. Likewise, the depth and width of molten pool expand with the laser radiation duration increasing. Meanwhile, the polished surface topography becomes smoother.

The use of teaching-learning based optimization technique for optimizing weld bead geometry as well as power consumption in additive manufacturing

Quality of weld bead geometry (width and height of metal deposited) and power consumption are still big challenges to the manufacture to control them in gas metal arc based additive manufacturing. The present study is aimed to optimize weld bead geometry and also made an attempt to reduce power consumption. Experiments are conducted at different torch angles, currents, wire feed speeds and welding speeds. Experimental results for width, height and depth of weld bead, power consumption and arc force are collected. Finite element method based numerical simulation is performed for width of molten pool to study its effect on the width and height of the weld bead. The process parameters are optimized using teaching-learning based optimization technique for achieving optimum weld bead geometry and power consumption. The proposed methodology found two optimal working conditions. Based on the power consumption, the optimal working condition-I is selected as best optimal working condition with optimum bead geometry such as 6.014 mm of width and 4.083 mm of height. The optimum power consumption is found to be 2496 W which is around 17%–41% less than that of experiments carried out. The optimal working condition is as follows: 124 A of current, 76.8° of torch angle, 8.38 m/min of wire feed speed and 0.42 m/min of welding speed.

Unidirectional Crystal Orientation of Dual-Phase Ni3Al-Based Alloy via Laser Irradiation

Dual-phase Ni3Al-based alloys feature extensive applicability even under high-temperature conditions. We selectively modified the microstructure of a representative dual-phase Ni3Al-based alloy from equi-axed grains to unidirectional grains, using a kW-class high-power laser irradiation technique. On employing the laser probe to linearly scan the Ni-9 at.% Al-16 at.% V alloy specimen, the laser-irradiated region was partially molten and then immediately solidified from the two edges of the molten pool toward the center. Laser irradiation under low-speed scanning increased the molten pool width. The grains in the laser-irradiated region extended preferentially from the two edges toward the center; their crystal orientation was similar to adjacent substrate grains, suggesting epitaxial growth. Therefore, the unidirectional orientation of grains could be extended via wide-range scanning using laser irradiation. This suggests that the microstructure of the alloy can be selectively modified to unidirectional orientated grains by optimizing laser irradiation conditions, such as the power density, scanning speed, and scanning paths. The hardness of the laser-irradiated region decreased due to the phase transformation from the ordered Ni3Al and Ni3V phases to the disordered fcc phase. However, the hardness improved to a value comparable to that of the alloy subjected to heat treatment at 980 °C for 1 h.

Modeling the influence of injection parameters on powder efficiency in laser cladding

For lateral powder injection, adjustments of defocusing distance, powder feeding angle, and molten pool width are set as input variables, and the powder transport vector, powder outflow position vector, and the injection offset vector are determined. The powder transport model involving above adjustment parameters is constructed based on operation of such vectors, which have been verified in experiments. Then, powder transport ratios of three variables with 13 × 13 × 5 levels are calculated to reflect the effect of injection parameters, which increases with the molten pool width, and maintains a high level as the defocusing distance and powder feeding angle are both adjusted positively (clockwise) or negatively (anti-clockwise). In addition, the powder transport ratio corresponding to an arbitrary combination of injection parameters can be calculated and its process window could be extended to determine a better process combination.

Virtual binocular vision sensing and control of molten pool width for gas metal arc additive manufactured thin-walled components

Gas metal arc (GMA) additive manufacturing (AM) is one of the significant wire and arc AM processes with the ability to produce large-scale metal parts in a layer by layer fashion. Despite this fact, techniques to realize process sensing and geometry control have not been perfectly developed. This study aims at molten pool width control in GMA AM using a passive vision sensing technique. A virtual binocular vision sensing system consisting of a biprism and a camera is designed to monitor the molten pool geometry. The molten pool width in a captured image pair is extracted by a series of procedures, such as sensor calibration, image pair rectification, disparity calculation, and width reconstruction. A verification test is conducted and reveals that the detection error of the sensing system is less than 3%. To keep consistent layer width in each layer, the deviation of the molten pool width is compensated by designing a fuzzy intelligent controller to adjust the arc current in real time. The effectiveness of the controller is evaluated via the deposition of thin-walled parts. The results indicate that the consistency of the molten pool width in GMA AM can be improved when employing the fuzzy controller.

Laser hot-wire cladding of Co-Cr-W metal cored wire

Laser hot-wire cladding (LHWC) is a hybrid deposition process where the wire is preheated by resistant heat during the deposition process. This process can dramatically increase the deposition efficiency and the material utilization rate. In this study, a cobalt-based metal cored wire was deposited by using laser hot-wire cladding. The CMOS camera assisted by a green laser illumination was used to monitor the stability of the deposition process and to investigate the molten pool width and length. The effects of processing parameters such as wire feed rate and scanning speed on the clad geometrical characteristics (height, wetting angle, and dilution rate) were studied in detail. The variation of the secondary dendrite arm spacing (SDAS) and microhardness with different processing parameters was discussed. The microstructure and corrosion resistance of the deposited alloy were finally analyzed. It was found that the resistant heat applied on the wire was a dominate factor influencing the stability of the deposition process. The clad height could be predicted by measuring the molten pool width. The molten pool length could be used as an indicator to reflect the cooling rate of the process. The clad height, wetting angle, and dilution rate were all sensitive to the wire feed rate and the scanning speed. The microhardness was mainly determined by the dilution rate. A high dilution rate decreased the clad hardness. The microstructure of the deposited alloy had a typical hypo-eutectic structure. An improved corrosion resistance of the deposited coating was obtained when compared to the substrate.