Numerical Simulation of Pool Surface Behavior in Full-Penetration TIG Welding

Literature Overview and Research Context

This paper by He Jingshan, Liu Wei, Zhang Binggang, and Wu Qingsheng from the State Key Laboratory of Advanced Welding Production Technology at Harbin Institute of Technology was published in the Welding Journal in 2007. The work addresses a fundamental challenge in TIG welding process modeling: accurately predicting the dynamic behavior of the molten pool surface during full-penetration welding operations. Full-penetration TIG welding is widely employed in thin-wall fabrication, aerospace structural components, and pressure vessel thin-section repairs, where the entire weld cross-section is melted and resolidified in a single pass. The pool surface behavior — including its shape, depression depth, oscillation patterns, and solidification dynamics — directly governs weld geometry, solidification defect formation, and final mechanical performance. This study is particularly relevant to engineers working in cladding and overlay welding, where pool geometry and surface tension effects at the molten pool boundary influence dilution rates and bond quality.

Core Technical Content

The authors developed a three-dimensional numerical model to simulate the pool surface behavior under the TIG arc, incorporating the full-penetration condition where the molten pool extends through the entire plate thickness. The model accounts for multiple physical phenomena simultaneously:

• Electromagnetic forces: Lorentz forces generated by the interaction of the welding current with the magnetic field, which drive convective flow within the pool and influence surface depression.
• Surface tension gradients: Marangoni convection driven by temperature-dependent surface tension coefficients, which is the dominant driving force for pool surface flow in TIG welding.
• Arc pressure: The mechanical pressure exerted by the plasma arc on the pool surface, which creates the characteristic arc crater depression.
• Buoyancy forces: Density variations due to temperature gradients contributing to natural convection.

The pool surface is modeled as a free boundary, meaning its position is not fixed but determined by the balance of forces acting upon it. This is a significant computational challenge because the surface shape evolves dynamically during the welding process, and the boundary conditions must be updated iteratively at each time step.

Key Technical Parameters and Modeling Approach

The model employs a moving coordinate system that follows the welding torch, transforming the moving boundary problem into a quasi-steady-state problem. This approach reduces computational cost while maintaining accuracy for continuous welding operations. The free surface is tracked using a level-set or front-tracking method, allowing the simulation to capture pool surface oscillations and the formation of the characteristic weld bead profile.

Interpretation of Key Findings

The simulation results reveal several important phenomena that are consistent with experimental observations but provide quantitative insights that are difficult to obtain through physical experiments alone. The pool surface depression is primarily governed by the competition between arc pressure pushing downward and surface tension pulling the surface inward. At higher current densities, the arc pressure dominates, resulting in deeper depressions and potentially narrower weld beads. The Marangoni convection pattern shows a strong dependence on the surface tension coefficient's temperature gradient: a negative gradient (surface tension decreasing with temperature) drives outward flow from the center to the edges, creating a shallow and wide pool, while a positive gradient produces inward flow, resulting in a deeper and narrower pool.

For full-penetration welding, the simulation captures the formation of the bottom weld bead and the transition from the front arc crater to the trailing solidification zone. The pool surface at the back of the arc exhibits a convex shape due to the balance of solidification shrinkage and surface tension, which is critical for predicting undercuts and incomplete penetration defects.

Relevance to Cladding and Overlay Welding Practice

Although the study focuses on full-penetration welding rather than overlay welding specifically, the fundamental pool dynamics are directly applicable to cladding operations. In TIG overlay welding for bimetallic products, the pool surface behavior determines the dilution rate between the cladding material and the base metal. A deeper pool depression implies greater penetration into the base metal and higher dilution, which can compromise the corrosion resistance or wear resistance of the overlay layer. Engineers working on nickel-based alloy cladding of carbon steel pressure vessels must carefully control the pool geometry to maintain dilution below acceptable limits — typically less than 5–10% for critical applications.

The simulation methodology also provides a framework for optimizing TIG overlay parameters. By varying the welding current, travel speed, and arc length in the numerical model, engineers can predict the resulting pool geometry and dilution without conducting extensive trial welds. This is particularly valuable for expensive overlay materials such as Inconel 625 or Hastelloy C276, where material cost is a significant consideration.

Key Questions and Reflections

Several questions arise from studying this work that are relevant to ongoing engineering practice. First, the accuracy of the surface tension coefficient and its temperature dependence is critical to the simulation results, yet these properties are difficult to measure accurately for many engineering alloys, especially at the high temperatures present in the weld pool. Second, the model assumes a single-phase pool, but in practice, the presence of flux, oxide films, and gas shielding variations can significantly alter the surface tension behavior. Third, for multi-pass overlay welding, the interaction between successive passes — including thermal history effects on the previously deposited cladding layers — is not captured in a single-pass simulation.

The study also raises the question of how well numerical predictions translate to industrial practice. While the model provides excellent insight into the fundamental physics, real welding operations involve uncertainties in arc characteristics, material property variations, and environmental factors that are difficult to incorporate into a deterministic model. A probabilistic approach or sensitivity analysis would provide more robust guidance for engineering decisions.

Study Insights and Implications

This research represents a mature application of computational fluid dynamics and heat transfer modeling to welding process optimization. The methodology established here has since been extended to include multi-physics coupling with electromagnetic field calculations, solidification modeling, and residual stress prediction. For engineers involved in bimetallic pressure vessel fabrication, the key takeaway is that pool surface behavior is not merely an academic curiosity but a practical determinant of weld quality, dilution control, and ultimately, the service life of the cladding layer. Understanding and controlling the pool surface through appropriate parameter selection is essential for producing reliable overlay welds that meet the demanding requirements of pressure vessel codes such as ASME VIII Division 1 and GB/T 150.

The work underscores the value of numerical simulation as a complementary tool to experimental welding trials. While physical experiments remain indispensable for validating process specifications and qualifying welding procedures, numerical models provide a cost-effective means of exploring the parameter space and identifying optimal welding conditions before committing to full-scale fabrication. Engineers who invest in developing proficiency with such simulation tools will be better equipped to tackle the increasingly complex challenges of modern cladding and overlay welding applications.