Finite Element Simulation of CLAM Steel TIG Welding: Temperature Fields and Molten Pool Dynamics
Why Simulation for Low-Activation Steels
CLAM steel (China Low Activation Martensitic steel) is a reduced-activation ferritic-martensitic steel developed for fusion reactor applications. It is a 9Cr-1W-0.2V-0.15Ta steel, designed to minimise the production of long-lived radioactive isotopes under neutron irradiation. Welding is a critical manufacturing step for fusion components, but the steel is sensitive to welding thermal cycles. Excessive heat input can cause grain growth in the heat-affected zone, reducing toughness. Too little heat input can cause incomplete fusion or cold cracking. This paper by Lei Yucheng and colleagues at Jiangsu University uses ANSYS finite element analysis to simulate the transient temperature field during single-side, double-pass TIG welding, with the goal of understanding how welding parameters affect the thermal cycle and the molten pool geometry.
Model Setup and Methodology
The model is a three-dimensional transient thermal analysis using the APDL parametric language and the element birth-death technique. The birth-death method is essential for simulating the gradual filling of the weld metal. As the heat source moves along the joint, elements in the weld zone are activated in sequence, representing the deposition of molten filler metal. The heat source is modelled as a moving Gaussian distribution, and the simulation includes preheating, which is standard practice for CLAM steel to prevent hydrogen-induced cracking.
The model was verified using infrared thermometry, which is a critical step. A simulation that has not been validated against experimental data is just a pretty picture. The authors compared the simulated temperature profiles at several locations with infrared measurements and found good agreement.
Results and Verification
The simulation provides quantitative data on the temperature distribution at different positions in the workpiece.
| Position | Peak Temperature | Cooling Rate (800-500°C) |
|---|---|---|
| Weld centerline | ~1800°C | Fast |
| HAZ (1 mm from fusion line) | ~1200°C | Moderate |
| Base metal (5 mm from fusion line) | ~700°C | Slow |
The results show that the cooling rate in the HAZ is fast enough to form fresh martensite, which is expected for a 9Cr steel. The simulation also examined the effect of welding speed on the molten pool shape. Increasing the welding speed from a baseline value to a higher value reduces the molten pool length and width, which is intuitive but important to quantify because it affects the risk of lack of fusion.
Practical Implications for Welding Procedure Development
For a fusion reactor component, the welding procedure must be qualified to very strict standards. The simulation approach described in this paper is a valuable tool for procedure development because it allows the engineer to explore the parameter space without welding dozens of test coupons. The key outputs that matter for procedure qualification are:
- The peak temperature in the HAZ, which determines the extent of grain growth.
- The cooling time from 800°C to 500°C (t8/5), which determines the final microstructure and hardness.
- The molten pool dimensions, which determine the risk of lack of fusion or burn-through.
In my experience, the t8/5 parameter is the most critical for 9Cr steels. For CLAM steel, the recommended t8/5 is typically in the range of 10-20 seconds. If the cooling time is too short, the HAZ is hard and susceptible to cold cracking. If it is too long, the HAZ is soft and the creep strength is reduced. The simulation allows the engineer to adjust welding parameters to hit the target t8/5 before ever striking an arc.
Reflections
This paper is a good example of how simulation can reduce cost and risk in welding procedure development. The fusion reactor industry is particularly suited to this approach because the components are expensive, the materials are difficult to weld, and the qualification requirements are stringent. However, the same approach can be applied to any high-value pipe or component. I have used finite element simulation to optimise welding procedures for thick-wall P91 pipe, where the t8/5 window is narrow and the cost of a failed procedure qualification is significant. The simulation allowed us to narrow the parameter window before welding the actual test coupons, saving both time and material.
The limitation of this approach is that the simulation is only as good as the material properties and heat source model used as input. The authors used temperature-dependent material properties and a Gaussian heat source, which are reasonable but not perfect. For critical applications, I would recommend a two-stage approach: first, run the simulation to narrow the parameter window, then validate with a small number of actual welding trials. That combination is the most efficient way to develop a robust welding procedure.
Zhuojin Pipe Fitting Co., Ltd