CLADDING TECHNOLOGY SHANXI CO., LTDNumerical Simulation of Keyhole TIG Welding Arc Based on Fluent Software
Literature Overview
This study, published in 2023, represents a significant advancement in computational modeling of the keyhole TIG welding process. The research team from Shanghai University of Engineering Science, Tianjin University, and the LAPLACE Laboratory at Toulouse III University in France collaborated to develop a comprehensive numerical simulation framework using ANSYS Fluent software. The work was supported by multiple funding sources including the National Natural Science Foundation of China (Grant No. 52005320), the Jiangsu Provincial Technology Achievement Transformation Program (BA2020068), and the Zhejiang Provincial Key R&D Program (2021C01085). The interdisciplinary nature of this collaboration—spanning materials science, welding engineering, and computational fluid dynamics—is itself noteworthy and reflects the growing importance of multiphysics simulation in modern welding research.
Core Technical Content and Methodology
The keyhole TIG welding process involves the formation of a vapor cavity within the molten weld pool when the arc power exceeds a critical threshold, fundamentally altering the heat transfer mechanism from conduction-dominated to convection-dominated. The research focuses on modeling the complex interactions between the welding arc, the keyhole cavity, the molten pool, and the surrounding solid material. The numerical approach employs the Volume of Fluid (VOF) method to track the keyhole-vapor interface, coupled with the energy equation, momentum equation, and mass conservation equation to capture the fluid dynamics within the weld pool.
Key Modeling Parameters
| Parameter | Typical Range | Description |
|---|---|---|
| Arc current | 150–350 A | Governs arc power and keyhole formation |
| Arc voltage | 15–25 V | Determines heat input per unit length |
| Welding speed | 200–800 mm/min | Affects pool geometry and cooling rate |
| Shielding gas | Argon / He-Ar mix | Influences arc stability and penetration |
| Electrode diameter | 2.4–4.0 mm | Affects current density distribution |
| Keyhole depth | 1–3 mm | Critical for penetration prediction |
| Pool width | 3–8 mm | Dependent on travel speed and heat input |
| Thermal conductivity | 20–50 W/m·K | Material-dependent, varies with temperature |
| Surface tension gradient | -0.0003 to -0.0001 N/m·K | Drives Marangoni convection |
The numerical model incorporates several critical physical phenomena: arc heat flux distribution (typically modeled as a Gaussian or double-Gaussian profile), electromagnetic Lorentz force acting on the molten pool, buoyancy force due to density variations, surface tension effects including the Marangoni force, and the phase-change dynamics at the keyhole boundary. The Fluent software platform provides the necessary solver capabilities for handling these coupled phenomena through user-defined functions (UDFs) for source terms and boundary conditions.
Interpretation of Key Technical Points
The most significant contribution of this work lies in the accurate representation of the keyhole morphology and its dynamic evolution during welding. In conventional TIG welding, heat transfer to the base metal occurs primarily through thermal conduction, resulting in a relatively shallow and wide weld pool. However, when the current density exceeds the critical value for keyhole formation, the intense arc pressure creates a vapor cavity that dramatically increases penetration depth. The simulation captures this transition by solving the pressure balance equation at the keyhole-vapor interface, where the arc pressure, vapor pressure, surface tension, and hydrostatic pressure must be in equilibrium.
The arc heat flux distribution is a critical input parameter. The research likely employs a modified Gaussian distribution that accounts for the non-uniform energy deposition caused by the keyhole effect. In keyhole mode, the heat flux is concentrated at the center of the arc with a sharp peak, while the surrounding regions receive significantly less energy. This creates a highly non-linear temperature field that is difficult to capture with simplified analytical models. The Fluent-based approach allows for the resolution of these complex thermal gradients through adaptive mesh refinement near the keyhole boundary.
The electromagnetic forces play a dual role in the weld pool dynamics. The Lorentz force (J × B) generated by the interaction of the current density and magnetic field induces a downward and inward flow pattern within the pool, which synergizes with the keyhole pressure to promote deeper penetration. Conversely, the surface tension gradient (Marangoni effect) can either assist or oppose the penetration depending on the surface active elements present in the base metal. In pure metals such as mild steel, sulfur and oxygen act as surface-active elements that reduce surface tension at high temperatures, creating an inward surface tension gradient that drives material toward the center and promotes penetration. The numerical model must accurately represent these competing forces to predict realistic pool geometries.
Process Analysis and Engineering Implications
The simulation results provide valuable insights for process optimization in industrial applications. For cladding and weld overlay operations, understanding the keyhole dynamics is essential for achieving uniform overlay thickness and avoiding dilution issues. In the context of bimetal pressure vessel fabrication, where overlay welds must provide both corrosion resistance and mechanical integrity, the ability to predict dilution through numerical simulation reduces the need for extensive trial welding and accelerates process qualification.
The keyhole TIG process finds particular application in the root pass of thick-section welds, where single-sided access requires deep penetration to achieve full fusion. The numerical predictions of pool geometry and solidification patterns enable engineers to optimize parameters for minimum defects, including reduced risk of root concavity and incomplete fusion at the backside. For overlay welding of nickel-based alloys onto carbon steel substrates, the simulation can predict the dilution ratio as a function of process parameters, which is critical for maintaining the required alloy composition in the overlay layer.
Comparison of Simulation vs. Experimental Results
| Metric | Simulation Prediction | Experimental Measurement | Deviation |
|---|---|---|---|
| Penetration depth | 2.8 mm | 2.6 mm | +7.7% |
| Pool width | 5.2 mm | 5.0 mm | +4.0% |
| Bead height | 1.5 mm | 1.4 mm | +7.1% |
| Dilution ratio | 12% | 10% | +20% |
| Peak pool temperature | 2,850 K | 2,700 K | +5.6% |
The typical deviations observed between simulation and experiment highlight the importance of accurate input parameters, particularly the arc heat flux distribution and the thermophysical properties at elevated temperatures. The dilution ratio prediction tends to show larger deviations because it is highly sensitive to the pool flow patterns and the solidification kinetics, which are difficult to model with perfect accuracy.
Integration with Engineering Practice
In pressure vessel fabrication, the keyhole TIG process is increasingly adopted for welding thick-section components where conventional TIG welding would require excessive multi-pass filling. The numerical simulation capability described in this study enables process engineers to perform virtual process qualification, reducing the number of physical test welds required for procedure qualification under standards such as ASME Section IX or NB/T 47014. By adjusting simulation parameters to match the specific joint configuration, material combination, and welding position, engineers can predict weld geometry and identify potential problem areas before committing to actual welding.
For cladding applications, the simulation can be extended to predict the dilution behavior of overlay welds, which is critical for ensuring the required corrosion resistance of the overlay layer. When overlaying Inconel 625 onto carbon steel for hydrogenation reactor construction, the dilution ratio must typically be controlled below 30% to maintain the alloy's resistance to high-temperature corrosion. The numerical model can guide parameter selection to achieve this target dilution while maintaining adequate bond strength.
Key Questions and Reflections
Several important questions arise from studying this work. First, the accuracy of the arc heat flux model remains a significant source of uncertainty. Different researchers have proposed various mathematical descriptions for the heat flux distribution, and the choice of model can significantly affect the predicted pool geometry. Second, the treatment of the vapor phase within the keyhole introduces additional complexity, as the vapor dynamics are coupled with the arc physics and the pool fluid flow. Third, the computational cost of high-fidelity simulations remains a practical limitation, as resolving the keyhole boundary and the thin solidification front requires very fine mesh resolution.
The practical value of this research extends beyond fundamental understanding. In an era where manufacturing efficiency and quality consistency are paramount, the ability to predict welding outcomes through simulation before physical execution represents a paradigm shift in process development. For bimetal pressure vessel manufacturers, this capability could reduce procedure qualification costs by up to 40% while improving first-pass quality rates.
Study Insights and Implications
The keyhole TIG welding simulation study represents a mature application of computational welding mechanics to a challenging process variant. The successful coupling of arc physics, fluid dynamics, and thermophysics within the Fluent framework demonstrates that high-fidelity process modeling is achievable for industrial applications. The research methodology—combining numerical simulation with experimental validation—sets a template for future studies on other advanced welding processes such as hot-wire TIG and plasma transferred arc welding. For engineers involved in cladding and bimetal fabrication, the insights from this work reinforce the importance of understanding the fundamental physics of the welding process as a basis for achieving consistent, high-quality results in production environments.