Weld Quality Research Articles

Classification of battery laser welding defects via enhanced image preprocessing methods and explainable artificial intelligence-based verification

This study focuses on the precise classification of normal and defective laser welding based on image data. A deep-learning algorithm with data augmentation and transfer learning is employed to enhance the classification accuracy. A comprehensive comparison is performed to evaluate the traditional preprocessing methods against the newly proposed and fused methods across three categories to discern their impact on the classification. Welding defects in the automotive battery industry pose significant cost issues during practical operation, necessitating compensation for quality problems. These issues raise concerns regarding trust and business relationships. Our research implications include substantial cost reduction and safety enhancement in the automotive battery industry by facilitating faster and more accurate laser welding quality assessments. To address the “black-box” nature of deep-learning algorithms, we use gradient-weighted class activation mapping (Grad-CAM) as an explainable artificial intelligence technique that provides intuitive validation and quantitative criteria for classification. We observe how different preprocessing methods influence the classification performance, identifying techniques that maximize accuracy. This study presents a straightforward yet highly accurate method for evaluating laser welding quality using only image data, achieving an accuracy of up to 98%. This value surpasses the accuracy obtained using raw data by 45%. Thus, our method serves as an effective tool for defect assessment across industries, thereby enhancing overall quality management. The innovative and impactful nature of this study underscores its significance for advancing the quality assessment of laser welding.

Investigation on the modeling and simulation of hydrodynamics in asymmetric conduction laser micro-welding of austenitic stainless steel and its process optimization

Laser micro-welding is a joining technology utilized across various high-value industries, like medical, automotive, e-mobility, and aerospace. A trial-and-error process to identify welding parameters does not necessarily lead to optimized quality levels. Furthermore, offline non-destructive examination methods often launched to verify welding quality may inadvertently trigger excessive costs and time delays, ultimately failing to guarantee defect-free welds. In response to these challenges, this article introduces an advanced multiscale model designed to unravel the intricate dynamics of hydrodynamics and the overarching physics within laser micro-welding melting pools. Developed using the COMSOL software package, the model adeptly demonstrates how surface tension gradients shape the geometry of welds, thus influencing their quality. This knowledge allows the mapping of welding defects. One of the novelties of the article is to introduce geometric dissimilar welding conditions by simulating an asymmetric edge joint. It shows a study on a new, unstudied way to laser weld with many applications in the field. The model further establishes its utility in design experiments to determine parameter, tolerance, and system design. Moreover, the insights garnered from understanding and controlling these drivers have far-reaching implications for the advancement of subsequent methodological research and the development of in-situ quality control practices by characterizing the welding defects. Finally, the results shows that the discouragingly high computational costs restrict its potential application to support a Digital Twin.

Development of a novel nonrigid support friction stir welding repair robot for aluminum alloy train

PurposeThis paper aims to develop a specific type of mobile nonrigid support friction stir welding (FSW) robot, which can adapt to aluminum alloy trucks for rapid online repair.Design/methodology/approachThe friction stir welding robot is designed to complete online repair according to the surface damage of large aluminum alloy trucks. A rotatable telescopic arm unit and a structure for a cutting board in the shape of a petal that was optimized by finite element analysis are designed to give enough top forging force for welding to address the issues of inadequate support and significant deformation in the repair process.FindingsThe experimental results indicate that the welding robot is capable of performing online surface repairs for large aluminum alloy trucks without rigid support on the backside, and the welding joint exhibits satisfactory performance.Practical implicationsCompared with other heavy-duty robotic arms and gantry-type friction stir welding robots, this robot can achieve online welding without disassembling the vehicle body, and it requires less axial force. This lays the foundation for the future promotion of lightweight equipment.Originality/valueThe designed friction stir welding robot is capable of performing online repairs without dismantling the aluminum alloy truck body, even in situations where sufficient upset force is unavailable. It ensures welding quality and exhibits high efficiency. This approach is considered novel in the field of lightweight online welding repairs, both domestically and internationally.

A Real-Time Monitoring Method for Droplet Transfer Frequency in Wire-Filled GTAW Based on Arc Sensing.

Droplet transfer frequency is a decisive factor in welding quality and efficiency in gas tungsten arc welding (GTAW). However, there still needs to be a monitoring method for droplet transfer frequency with high precision and good real-time performance. Therefore, a real-time monitoring method for droplet transfer frequency in wire-filled GTAW using arc sensing is proposed in this paper. An arc signal acquisition system is developed, and the wavelet filtering method filters out noise from the arc signal. An arc signal segmentation method-based on the OTSU algorithm and a feature extraction method for droplet transition based on density-based spatial clustering of applications with noise (DBSCAN)-is proposed to extract the feature signal of the droplet transition. A new conception of droplet transition uniformity is proposed, and it can be used to monitor the weld bead width uniformity. Numerous experiments for monitoring droplet transfer frequency in real time are conducted with typical welding parameters. This method enables the real-time observation of droplet transfer frequency, and the result shows that the average monitoring error is less than 0.05 Hz.

Dynamic wetting and spreading mechanisms of SAC305 with added Au elements in the laser jet weld ball bonding process

The lead-free solder SAC305 has inferior wetting properties compared to conventional leaded solders, which deteriorates the weld quality. In this study, Au coatings with varying thickness were applied to improve its wettability and the influencing mechanism was investigated. The results indicated that the wetting angle (WA) and spreading area (SA) increased and then decreased with increasing thickness of the Au coating. The minimum value of WA and maximum value of SA were 25.4° and 8.7 mm2, respectively, when Au coating thickness was 50 nm. The analysis of the interfacial microstructure revealed that the generation of AuSn4 in the triple line region would significantly improve the wettability of Sn alloys. However, when the Au coating thickness reach up to 75 nm, the dense intermetallic compound (IMC) (Cu, Ni)6Sn5 was generated and the wettability of Sn alloys was reduced. The effect of the interfacial compounds on wettability depended on the change of energy consumption mechanism during wetting. The main energy consumption mechanism changed from adsorption of elements to a reaction-controlled, which reaction wetting dominated when the Au coating thickness was greater than 50 nm. This study will provide a deep understanding of wetting behavior and a valuable insight for optimizing solder wetting processes.

Influence of oscillation parameters on energy distributions and dynamic behaviors during laser welding of aluminum alloy T-joints assisted with solder patch

The laser energy distribution on base material (BM) surface plays an important role in the welding quality. To reduce pore and spatter welding defects during laser welding of aluminum alloy T-joints, a circular shaped oscillating laser beam welding (OLBW) method for aluminum alloy T-joints assisted with solder patch is proposed in this paper. An energy distribution model and a dynamic model for OLBW of aluminum alloy T-joints assisted with solder patch are established to analyze the influence of oscillation parameters on energy distributions of BM surface and dynamic behaviors of molten pool and keyhole. The energy distribution characteristics are obtained by the energy distribution model and the dynamic behaviors of molten pool and keyhole are calculated by the dynamic model for OLBW of aluminum alloy T-joints assisted with solder patch. The results show that the uniformity of energy distribution is improved with the increase in oscillation amplitude or frequency. The oscillation parameters can be adopted to adjust the keyhole stability and the temperature distributions, flow velocities and morphologies of molten pool, which is of great importance for aluminum alloy T-joints welding with high quality.

A new method for deep learning detection of defects in X-ray images of pressure vessel welds.

Given that defect detection in weld X-ray images is a critical aspect of pressure vessel manufacturing and inspection, accurate differentiation of the type, distribution, number, and area of defects in the images serves as the foundation for judging weld quality, and the segmentation method of defects in digital X-ray images is the core technology for differentiating defects. Based on the publicly available weld seam dataset GDX-ray, this paper proposes a complete technique for fault segmentation in X-ray pictures of pressure vessel welds. The key works are as follows: (1) To address the problem of a lack of defect samples and imbalanced distribution inside GDX-ray, a DA-DCGAN based on a two-channel attention mechanism is devised to increase sample data. (2) A convolutional block attention mechanism is incorporated into the coding layer to boost the accuracy of small-scale defect identification. The proposed MAU-Net defect semantic segmentation network uses multi-scale even convolution to enhance large-scale features. The proposed method can mask electrostatic interference and non-defect-class parts in the actual weld X-ray images, achieve an average segmentation accuracy of 84.75% for the GDX-ray dataset, segment and accurately rate the valid defects with a correct rating rate of 95%, and thus realize practical value in engineering.

Microstructure and Mechanical Properties of Powder Metallurgy Superalloy Joints Welded by Inertia Friction Welding.

In recent years, for the structural characteristics and design requirements of the integral rotor and disc shaft of the integrated engine, the welding quality and mechanical properties of superalloy weldments have received increasing attention. In this paper, inertia friction welding (IFW) of FGH96 alloy was carried out using different welding parameters, and the homogeneous connection of FGH96 alloy hollow bars was successfully realized. The microstructure evolution, mechanical properties and fracture failure of the welded joints at room and high temperatures were investigated. The FGH96 alloy IFW joints were divided into the weld nugget zone (WNZ), the thermo-mechanically affected zone (TMAZ), the heat-affected zone (HAZ) and the base metal (BM), and there were significant differences in grain structure and distribution of the γ' phase in each of the characteristic zones. The microhardness and tensile properties of the IFW joints were investigated, and the results showed an "M"-shaped curve in the microhardness distribution, with the lowest point of hardness observed in the HAZ. The tensile test results indicated that the fracture position moved from the BM to the WNZ with the increase in temperature, the microstructure at the fracture changed significantly and the tensile strength decreased from 1512.0 MPa at room temperature to 1201.3 MPa at 750 °C. The difference in the mechanical properties of the joints was mainly attributed to the changes in the dissolution and precipitation of the γ' phase.

Numeric simulation and experimental study in magnetic field-assisted laser scanning welding of T2 red copper

To address the challenges of low energy coupling efficiency and poor process stability in the laser welding of T2 red copper, this study establishes a thermal-fluid-coupled three-dimensional numerical model; and we conduct experimental tests to investigate the effects of different magnetic field strengths on laser scanning welding of T2 red copper. The results indicate a high degree of conformity between the simulated calculations and experimental tests in terms of the cross-sectional fusion morphology. As the magnetic field strength increases from 0 to 160 mT, the lowest and highest temperatures, melt width, and melt depth of the weld pool initially exhibit an initially increasing trend followed by a decreasing trend. At a magnetic field strength of 120 mT, these parameters peak at 1383 K, 4810 K, 2.388, and 1.753 mm, representing improvements of 2.44 %, 58.75 %, and 1.68 %, respectively, compared with those without an applied magnetic field. However, the maximum flow velocity within the weld pool exhibits a trend of initially decreasing and subsequently increasing, reaching a minimum value of 0.4102 m/s at a magnetic field strength of 120 mT, representing a 12.76 % reduction compared with that without an applied magnetic field. As the magnetic field strength increases, the surface morphology of the weld seam initially improves and then deteriorates. In addition, the grain structure within the weld zone first becomes finer and coarser. When the magnetic field strength reaches 120 mT, the surface-forming quality of the weld seam improves; the internal grain structure of the weld zone becomes finer; and the hardness, tensile strength, and electrical conductivity of the weld joint reach their respective maximum values of 47.13 HV, 270 MPa, and 78.24 mS/m. This represents significant improvements of 18.45 %, 56.07 %, and 25.2 %, respectively, compared to those without an applied magnetic field. Under a steady magnetic field. the effective thermal input of the laser energy to the base material is enhanced, thereby improving the stability of the weld pool during the welding process. This enhancement leads to improved welding quality, yielding new insights and theoretical foundations for the laser welding of T2 red copper.

An intelligent inspection method for body-in-white weld quality based on vibration excitation response signals

To enhance the assembly quality in Body-in-White (BIW) assembly, this paper proposes an intelligent detection method for the nugget quality of Resistance Spot Weld (RSW) based on weld joint vibration excitation response signals. The method proposes a novel deep learning model, the Local-Global Hierarchical Graph Neural Network (LGHGNN). LGHGNN can automatically construct graph structures and, by introducing a newly designed upgrade pooling operation, extends the traditional flat structure of graph networks into a hierarchical structure within three-dimensional space. Therefore, LGHGNN achieves layered interaction of local-global information, enabling the model to focus on local details while gaining a broader learning perspective. Additionally, this paper proposes a strategy for multi-label unsupervised anomaly detection that involves layered interaction and collaborative decision-making for local and global graphs. The effectiveness of LGHGNN is demonstrated through its application in the BIW right front door assembly, achieving a remarkable 97.5% average accuracy in multi-region parallel anomaly detection.

Development of a novel fabricating thin-walled TA2 titanium tube via high-frequency induction welding

Due to the problems of low welding efficiency, large heat-affected zone, and poor welding quality in the process of welding thin-walled titanium tubes by argon arc welding, there are few studies on the use of high-frequency induction welding (HFIW) of thin-walled titanium alloy tubes. The evolution law of weld microstructure and mechanical properties of the thin-walled titanium tube needs to be further studied because of rapid welding speed and the small heat-affected zone of HFIW. Therefore, a novel manufacturing method via high-frequency induction welding is proposed in this paper to solve the existing problems. With an industrial-grade titanium TA2 tube (wall's thickness is 0.5 ​mm) as the research object, a comparative study is conducted in this research to examine the morphology, microstructure, microhardness, and tensile characteristics of welded joints at different welding power. The findings demonstrated a significant efficacy of HFIW in resolving these challenges. The mechanical properties and microstructue of heat-affected zone (HAZ) were characterized. The lowest hardness is measured at 202 HV, while the base material was recorded as 184 HV, when the welding speed of HFIW is set at 50 ​m/min. Meanwhile, the heat-affected zone has the highest hardness at 224 HV, a tensile strength of 446.8 ​MPa and a post-fracture elongation of 16%. The results showed that HFIW can not only greatly improve the welding efficiency, significantly improve the microstructure of weld joint and HAZ, and improve the mechanical properties of thin-walled titanium pipe, but also provide a highly feasible welding method for welding ultra-thin-walled pipes.

Structural design and adaptive tracking control of automatic welding robot for liquefied natural gas containment system

Liquefied natural gas (LNG) is commonly transported by LNG carriers and stored in cargo containment systems. The primary barrier of the MARK III cargo containment system is welded to a closed space with corrugated stainless steel plates. To meet the requirements of excellent sealing and thermal insulation for cargo containment, the welding process of the corrugated plate need to be strictly controlled, which poses a challenge to the development of related welding equipment. In this paper, we present a new five-axis automatic welding robot system used for plasma arc welding on corrugated surfaces. The moment transfer scheme of the dual linkage mechanism makes the rotary movement of the robot more accurate and stable which gives the system with simpler control algorithm and better overall force characteristics. To ensure tracking accuracy, a novel sensing method based on a LASER sensor, two contact sensors, and an angle sensor is proposed to implement multiple functions such as tracking the welding arc length, identifying corrugation shapes, and detecting welding gun posture. Based on the devised tracking sensor system and welding robot, a servo-control system with a surface-tracking welding control algorithm is established. The experimental results show that the robot system's welding speed is about 7 times that of hand welding and the welding qualification rate was 99%, significantly improving welding efficiency and quality as a critical equipment technology in the process of corrugated plate welding.

Optimization of Installation Position for Complex Space Curve Weldments in Robotic Friction Stir Welding Based on Dynamic Dual Particle Swarm Optimization

Robotic friction stir welding (RFSW), with its wide application range, ample working space, and task flexibility, has emerged as a vital development in friction stir welding (FSW) technology. However, the low stiffness of serial industrial robots can lead to end-effector deviations and vibrations during FSW tasks, adversely affecting the weld quality. This paper proposes a dynamic dual particle swarm optimization (DDPSO) algorithm through a new comprehensive stability index that considers both the stiffness and vibration stability of the robot to optimize the installation position of complex space curve weldments, thereby enhancing the robot’s stability during the FSW process. The algorithm employs two independent particle swarms for exploration and exploitation tasks and dynamically adjusts task allocation and particle numbers based on current results to fully utilize computational resources and enhance search efficiency. Compared to the standard particle swarm optimization (PSO) algorithm, the DDPSO approach demonstrated superior search capabilities and stability of optimization results. The maximum fitness value improved by 4.2%, the average value increased by 12.74%, and the concentration level of optimization results rose by 72.91% on average. The new optimization method pioneers fresh perspectives for optimizing the stability of RFSW, providing significant grounds for the process optimization and offline programming of complex spatial curve weldments.

Development of a novel autogenous double pulse tungsten inert gas welding technique for joining IN625 in concentrated solar power plants

The research focuses on the challenges of joining the IN625 alloy for high-temperature applications. Tungsten inert gas (TIG) welding has drawbacks such as higher heat input (HI), which causes elemental segregation and secondary phases, ultimately affecting the quality of weldments. Previous literature indicates that single-pulsed TIG (SP-TIG) welding has shown improvements over TIG welding. In this study, a new technique called double-pulsed TIG (DP-TIG) welding was developed to address the HI and elemental segregation issues during the welding process. The study also compares the improvements in properties of DP-TIG with SP-TIG welding. The results demonstrate that DP-TIG significantly reduces HI by 53% compared to SP-TIG. It effectively decreased residual stress by 26% and suppressed secondary phase formation. Moreover, DP-TIG exhibits superior mechanical properties, with an 11% increase in hardness and a 10% increase in tensile strength compared to SP-TIG. These findings establish DP-TIG as the preferred method for joining the IN625 alloy.

3-D FE modelling of residual stress distribution in 3003 aluminium alloy overlapped joints by ARM laser with beam oscillations

Adjustable Ring Mode (ARM) laser technology combined with Beam Oscillation optics module was introduced to effectively enhance weld quality in joining aluminium alloys. This study developed a three-dimensional (3-D) thermo-mechanical finite element (FE) model to analyze the distribution of thermally induced residual stresses in laser-welded 3003 aluminium alloy. The accuracy of model was validated through comparisons with experimental weld cross-sections and further reinforced by X-ray diffraction residual stress measurements. The transient temperature field and stress distribution during the welding process were investigated using the verified model. Additionally, the hardness and microstructure of the welded samples were examined through micro-hardness tests and scanning electron microscopy. Results showed that welding with beam oscillation resulted in less penetration but higher hardness and finer columnar grain size in the fusion zone. Additionally, the influence of varying standoff distances between two sequential weld beads on residual stress distribution was investigated. Numerical findings revealed that higher residual stress concentrations were located at the weld seam and dominated by tensile longitudinal residual stress and compressive transverse residual stress. As the distance between weld beads increased, transverse and equivalent residual stresses decreased, while longitudinal residual stress away from the bead increased.

Control of meltpool shape in laser welding

In laser welding, the achievement of high productivity and precision is a relatively easy task; however, it is not always obvious how to achieve sound welds without defects. The localised laser energy promotes narrow meltpools with steep thermal gradients, additionally agitated by the vapour plume, which can potentially lead to many instabilities and defects. In the past years, there have been many techniques demonstrated on how to improve the quality and tolerance of laser welding, such as wobble welding or hybrid processes, but to utilise the full potential of lasers, we need to understand how to tailor the laser energy to meet the process and material requirements. Understanding and controlling the melt flow is one of the most important aspects in laser welding. In this work, the outcome of an extensive research programme focused on the understanding of meltpool dynamics and control of bead shape in laser welding is discussed. The results of instrumented experimentation, supported by computational fluid dynamic modelling, give insight into the fundamental aspects of meltpool formation, flow direction, feedstock melting and the likelihood of defect formation in the material upon laser interaction. The work contributes to a better understanding of the existing processes, as well as the development of a new range of process regimes with higher process stability, improved efficiency and higher productivity than standard laser welding. Several examples including ultra-stable keyhole welding and wobble welding and a highly efficient laser wire melting are demonstrated. In addition, the authors present a new welding process, derived from a new concept of the meltpool flow and shape control by dynamic beam shaping. The new process has proven to have many potential advantages in welding, cladding and repair applications.

A novel optimizing Pulsed Plasma Gas Variable Polarity Plasma Arc Welding (PPG-VPPAW) method for improving weld quality

To enhance the reliability of aluminum alloy welding quality, a novel Pulsed Plasma Gas Variable Polarity Plasma Arc Welding (PPG-VPPAW) system is proposed. This system enables independent control and rapid adjustment of the plasma gas, allowing for flexible modulation of arc pressure output under constant thermal conditions, thereby reducing porosity and refining grain structure. The research analyzed the impact of parameters of different plasma gas on arc characteristics, welding porosity, and welding microstructure. The results reveal that the periodic changes in pulsed plasma gas induce corresponding variations in arc temperature, arc shape, arc pressure, arc voltage, and the morphology of the weld pool keyhole. The magnitude of the difference between the base value and peak value of pulsed plasma gas determines the extent of arc pressure fluctuations and the behavior of the backside keyhole. The maximum oscillation amplitude of the backside keyhole can reach up to 4.5 mm². Additionally, the oscillation of pulsed plasma gas within the molten pool has been demonstrated to effectively reduce weld porosity and refine grain size. Under this process, the grain size of the weld metal can be reduced by up to approximately 25%. Current work provides an important direction for fabricating highly reliable aluminum alloy weldments.

Ultrasonic non-destructive evaluation study of molecular diffusion bonding of thin copper-aluminum electrode sheets

The weld quality of copper and aluminum thin electrode sheets in molecular diffusion bonding was non-destructively evaluated using ultrasonic resonance techniques. During the welding process, the intermediate layer material nickel diffuses into the molecules of both the copper sheet and aluminum sheet, resulting in the formation of a solid solution phase layer. This leads to a 5-layer structure in the welded body. If there are defects in this solid solution phase layer, it can cause mutations in the ultrasonic resonance signals within the weld body. In order to characterize the weld quality between copper and aluminum sheets, an acoustic attenuation coefficient was introduced. The ultrasonic resonance signals within the weld body of copper and aluminum thin electrode sheets were analyzed under four different welding states. Experimental testing revealed significant differences in acoustic attenuation coefficients among these different welding states. A smaller acoustic attenuation coefficient indicates better welding quality. Therefore, by setting a reasonable threshold for this coefficient, it is possible to effectively evaluate the welding quality of molecular diffusion bonding between copper and aluminum thin electrode sheets.

Modelling the Evolution of Phases during Laser Beam Welding of Stainless Steel with Low Transformation Temperature Combining Dilatometry Study and FEM

In this study, the evolution of volume fractions during laser beam welding (LBW) of stainless steel, with a specific focus on incorporating the low transformation temperature (LTT) effect using the dilatometer, has been proposed. The LTT effect refers to the phase transformations that occur at lower temperatures and lead to the formation of a martensitic microstructure, which will significantly influence the residual stresses and distortion of the welded joints. In this research, the LTT conditions are achieved by varying the Cr and Ni content in the weld seam by varying the weld parameter, including laser power, welding speed and filler wire speed. The dilatometer analysis technique is employed to simulate the thermal conditions encountered during LBW. By subjecting the stainless steel samples to controlled heating and cooling cycles, the kinetics of the volume fractions can be measured using the lever rule and empirical method (KOP and Lee). The phase transformation simulation model is computed by integrating the thermal and metallurgical effects to predict the volume fractions in LBW joints and has been validated using dilatometer results. This provides valuable insight into the relationship between welding parameters and phase transformations in stainless steel with the LTT effect during laser beam welding. Using this relationship, the weld quality can be improved by reducing the residual stresses and distortion.

Intelligent optimization of non-rigidly supported stationary shoulder friction stir welding parameters

During the online welding of railway aluminum alloy trucks using fixed shoulder stirring friction welding technology, defects such as flying edges, cracks, and tunnels may occasionally occur due to poor coordination of the welding process parameters. This inconsistency can compromise the success rate of the welding. To enhance the quality and efficiency of the welding repairs, an optimization of the welding process parameters of the developed equipment was undertaken. An intelligent prediction model for welding parameters was established based on the BSO-BP neural network. The model was trained with 17 sets of authentic welding data, and its accuracy was validated through a comparison of predictive outputs with actual welding outcomes. The BSO-BP neural network demonstrated a maximum error of 1.85% and a relative error of 1.27%, showing a marked improvement over traditional BP neural network predictions. Utilize the optimized welding process parameters obtained from a trained predictive model to conduct welding experiments under actual working conditions. During the experiment, the deformation of the test plates was minimal. The resulting welds were smooth, free of defects such as flash and grooves, and the overall welding quality was good. The tensile strength of the welds, at 262.95 MPa, was close to the predicted results and reached 78.49% of the base material’s tensile strength. In the microstructure of the weld, sporadically distributed dimples of varying sizes can be observed, indicating the presence of second phase particles which contribute to enhancing the tensile strength of the weld, thereby meeting the strength requirements of the aluminum alloy car body. The parameter optimization method used has certain reference value and can provide guidance for similar research.