Numerical Simulation of Arc and Molten Pool Interaction in Stationary TIG Welding: A Veteran's Reading of the Free Interface Problem
Literature Overview and Why the Free Interface Matters
This paper from the Welding Journal (焊接学报), published in 2005, comes from Shanghai Jiao Tong University. The authors, Lu Fenggui, Tang Xinhua, Li Shaoqing, Yao Shun, and Lou Songnian, set out to do something that sounds academic but is actually deeply practical: model the interaction between a TIG welding arc and the molten pool, including the deformation of the free surface. When I first read the abstract, I thought, "This is another simulation paper that will never make it to the shop floor." But as I read further, I realized that the free interface problem is exactly the issue that determines whether a weld looks good or looks like a mess.
In TIG welding, the arc is not a rigid column. It is a plasma that interacts with the molten pool, and the pool surface is not flat. The arc pressure pushes the pool down, creating a depression. The electromagnetic force and the surface tension try to restore the surface. The result is a constantly moving, deforming free interface. If you cannot model that, you cannot predict the weld penetration, the weld bead shape, or the occurrence of defects like undercut or humping. This paper is one of the early attempts in China to build a unified three-dimensional mathematical model that couples the arc and the pool through the free interface.
The Core Technical Problem: A Moving, Deforming Boundary
The authors correctly identify that the key challenge is the treatment of the free interface. In a conventional welding simulation, you often assume a flat pool surface or a simple geometry. But in reality, the interface between the arc plasma and the molten metal is a complex, moving boundary. The arc pressure, the Lorentz force, the surface tension, and the buoyancy force all act on this boundary. The authors propose a three-dimensional unified mathematical model that solves the arc equations and the pool equations simultaneously, with the free interface as a moving boundary condition.
This is a significant technical achievement, and I want to explain why it matters on the shop floor. When you weld a thick section with a TIG torch, the arc pressure can be high enough to create a keyhole or at least a significant depression. If the depression is too deep, the arc can penetrate through the pool and cause a burn-through or a lack of fusion on the side walls. If the depression is too shallow, you get a wide, shallow weld with poor penetration. The shape of the free interface determines the weld profile, and the weld profile determines whether you pass or fail the visual and radiographic inspection.
The authors validated their model against experimental measurements of the molten pool shape. They found that the calculated pool shape was in good agreement with the measured shape. This is important because it means the model can be used to predict the weld bead geometry for different welding parameters without running a full experimental matrix. That saves time and money, especially when you are qualifying a new procedure.
Interpretation of the Physical Mechanisms
The paper reveals the mechanism of arc-pool interaction in a way that is not obvious from experiments alone. The arc pressure is the dominant force that deforms the pool surface. As the arc pressure increases, the pool surface is pushed down, creating a depression. The electromagnetic force, which is caused by the current diverging in the pool, creates a circulation pattern that helps to transport heat from the arc to the pool bottom. The surface tension acts to restore the surface, but it is not strong enough to overcome the arc pressure at high currents.
The authors also show that the free interface is not static. It oscillates, and the oscillation frequency depends on the welding parameters. This is a critical insight. If the oscillation frequency matches the natural frequency of the pool, you can get resonance, which leads to unstable weld bead formation. This is a known problem in TIG welding, especially at high currents. The model can help you identify the conditions that lead to instability and avoid them.
From a practical standpoint, this means that the welding parameters that give you a stable weld in one position may not give you a stable weld in another position. The pool shape and the free interface are affected by gravity, so a horizontal weld will behave differently from a vertical or overhead weld. The model can be extended to account for gravity, but the authors did not do that in this paper. That is a limitation, but it does not reduce the value of the work.
Practical Implications for Procedure Qualification and Defect Prevention
Let me connect this to what we do in the shop. When we qualify a TIG welding procedure for a pipe or a fitting, we run a series of coupons at different parameters and then we test them. That is the traditional approach. It works, but it is expensive and time-consuming. The model described in this paper offers a way to reduce the number of trials. If you can predict the weld bead shape for a given set of parameters, you can narrow down the window before you ever strike an arc.
The model is also useful for understanding defects. Undercut, for example, is caused by the arc pressure pushing the pool surface down at the edges of the weld. If the arc pressure is too high, the pool surface is depressed too much, and the filler metal cannot fill the gap. The model can help you identify the conditions that lead to undercut and adjust the parameters accordingly. Similarly, lack of fusion is often caused by the arc not penetrating deep enough. The model can show you the penetration depth for different parameters, and you can adjust the current or the travel speed to ensure full penetration.
I have seen many cases where a welder uses a high current to get more penetration, but the result is a wide, shallow weld with undercut. The model would have shown that the arc pressure at high current creates a shallow depression, not a deep one, because the pool becomes wider and the pressure is distributed over a larger area. This is counterintuitive, but it is exactly the kind of insight that a good model can provide.
Key Questions and Reflections
- Is the model applicable to moving TIG welding? The authors studied a stationary arc, which is simpler. In practice, most TIG welding is done with a moving torch. The model would need to be extended to include the travel speed and the asymmetry of the pool. That is a natural next step, and I would expect the same principles to apply, but the free interface would be more complex.
- How does the model handle the transition from a depression to a keyhole? In a stationary TIG weld, the arc pressure can be high enough to create a keyhole, but the authors did not discuss this. In practice, a keyhole is usually associated with plasma welding or laser welding, not TIG. But at high currents, a TIG arc can create a deep depression that approaches keyhole conditions. The model should be able to capture this, but it would require a more sophisticated treatment of the free interface.
- What about the effect of shielding gas composition? The arc properties depend on the shielding gas. Argon gives a different arc pressure and arc temperature than helium or a mixture. The model does not explicitly include the gas composition, but it could be added by modifying the arc properties. This is important for our shop because we often use argon-helium mixtures for thick section welding.
- Can this model be used for real-time control? That is the ultimate goal. If you can measure the arc voltage or the arc light and feed that into the model, you could predict the weld bead shape in real time and adjust the parameters to maintain a consistent weld. This is still a research topic, but the foundation laid in this paper is a step in that direction.
Study Insights and Implications for Our Shop
This paper is a reminder that welding is not just an art. It is a science, and the physics of the arc and the pool can be modeled and predicted. The free interface problem is the key to understanding weld bead formation, and this paper provides a solid foundation for that understanding.
For our shop, the practical implication is that we should not rely solely on trial and error. We should use models like this to narrow down the parameter window before we run a procedure qualification. This saves time and money, and it gives us a better understanding of why a particular set of parameters works or does not work.
I would also note that the model is not a substitute for experience. The validation was done on a stationary arc, and the results may not directly apply to a moving torch. But the principles are sound, and the model can be extended. I would encourage our engineers to read this paper and to think about how the free interface affects the weld quality in our specific applications.
In the end, this paper is a good example of how academic research can inform industrial practice. It does not give us a turnkey solution, but it gives us a tool to think with. And in the welding business, thinking before you weld is always better than grinding out a defect after you weld.
Zhuojin Pipe Fitting Co., Ltd