Numerical Simulation of Temperature and Residual Stress Fields in IN738 TIG Butt Welding
Core Findings and Numerical Approach
This study by Nan Qing and colleagues at Xi'an Thermal Power Research Institute tackles a problem I have wrestled with on the shop floor many times: what happens to a nickel-based superalloy when you put a TIG torch on it, and how do you predict the residual stress state before you ever strike an arc. The team used MSC.Marc finite element analysis to model IN738 butt welding under different welding currents, and the results are worth reading carefully because they confirm what many of us have suspected but rarely quantified.
The key finding is that welding current has a direct, positive correlation with the maximum temperature reached, the molten pool size, and the magnitude of residual stresses. This is not surprising to anyone who has run a TIG torch, but the quantification is valuable. What caught my attention was the stress distribution pattern at the weld centerline surface: both transverse and longitudinal residual stresses are tensile. This is critical because for a high-temperature alloy like IN738, which is used in turbine blades and combustion chambers, tensile residual stress at the surface is exactly what you do not want when the component will see cyclic thermal loading in service.
Residual Stress Distribution Analysis
The study reports that at the mid-section of the welded component, the longitudinal residual stress is tensile in the weld zone, reaches a maximum in the heat-affected zone (HAZ), and transitions to compressive stress far from the weld. The transverse residual stress remains relatively small and is entirely tensile across the section. This pattern is consistent with what I have seen in actual welded components, and it explains why HAZ cracking is such a persistent problem in these alloys.
Let me put this in a table for clarity, because the distribution matters more than the absolute values when you are deciding whether to post-weld heat treat or apply mechanical stress relief.
| Location | Longitudinal Residual Stress | Transverse Residual Stress | Practical Concern |
|---|---|---|---|
| Weld centerline surface | Tensile | Tensile | Surface cracking risk under cyclic loading |
| Weld zone (mid-section) | Tensile | Small tensile | Acceptable if PWHT applied |
| HAZ (mid-section) | Maximum tensile | Small tensile | HAZ cracking risk, especially with high restraint |
| Far from weld (mid-section) | Compressive | Small tensile | Generally safe, but distortion may occur |
Practical Implications for High-Temperature Alloy Welding
Now let me tell you what this means for the plant floor. IN738 is not a forgiving alloy. It has high gamma-prime content, which makes it prone to strain-age cracking during post-weld heat treatment and to HAZ liquation cracking during welding. The simulation confirms that higher welding current increases residual stress, which means you cannot simply crank up the heat to get better penetration. There is a sweet spot, and you have to find it.
In my experience, the practical approach is to use a low heat input with a preheat of around 200-300°C, keep interpass temperature under control, and consider a controlled post-weld heat treatment schedule. The simulation gives you a tool to optimize these parameters before you ever weld a test coupon. That saves time, money, and scrap.
I also note that the study used MSC.Marc, which is a well-established code, but the accuracy of any welding simulation depends heavily on the material properties you input. For IN738, you need temperature-dependent thermal conductivity, specific heat, and mechanical properties, plus a good understanding of the phase transformations. If you do not have those, the simulation is just a fancy guess.
Lessons Learned and Engineering Insights
Here is where I want to be blunt. I have seen too many projects where the welding engineer runs a simulation, gets a pretty color contour plot, and then signs off on a procedure without validating it against actual weld trials. The simulation is a guide, not a substitute for experience. The residual stress pattern described in this paper is consistent with what I have observed in actual IN738 components, but the absolute values depend on boundary conditions, restraint, and the exact welding sequence.
The key takeaway for the shop floor is this: control the welding current, control the heat input, and do not assume that a higher current will give you a better weld. It will give you a bigger molten pool, a higher peak temperature, and more residual stress. For a high-temperature alloy that will see service at 700-900°C, that residual stress is a ticking time bomb. You need to manage it with proper preheat, interpass control, and post-weld heat treatment.
In practice, I would also recommend considering mechanical stress relief methods such as ultrasonic impact treatment or peening for critical components, because the tensile residual stress at the surface is the most dangerous. The simulation tells you where the stress is; your job is to do something about it.
Concluding Remarks
This paper is a solid contribution to the welding literature for high-temperature alloys. It confirms the fundamental relationship between welding current, heat input, and residual stress, and it provides a numerical framework that can be used to optimize welding parameters before committing to production. The practical value lies in using the simulation to establish a process window, then validating it with actual weld trials and residual stress measurement. For those of us who have spent decades welding nickel-based superalloys, the message is clear: respect the material, control the heat, and never ignore residual stress. It will find you in service if you do.
Zhuojin Pipe Fitting Co., Ltd