ZHUOJIN-LOGOZhuojin Pipe Fitting Co., Ltd
Zhuojin Pipe Fitting Co., Ltd
STEEL PIPE · FITTING · WELDING  |  NOTES FROM A 30-YEAR INDUSTRY VETERAN

Numerical Simulation of End Deformation in Thin-Walled Cylinders Under Different Overlay Welding Parameters

Overview and Context

This paper by Wu Hongshuang and colleagues addresses a common and costly problem in the fabrication of large-diameter thin-walled cylinders: the distortion of the cylinder ends after the internal overlay welding of stainless steel. This distortion makes the subsequent assembly difficult, and in severe cases, it can render the cylinder unusable. The authors use the Marc finite element software to simulate the welding temperature field and the resulting deformation, and they compare the effects of different welding sequences, welding speeds, end clearance, and number of overlay layers.

This is a highly practical problem. In my experience with the fabrication of large-diameter pipe, vessels, and separators, the distortion of thin-walled components during overlay welding is a major source of rework and schedule delays. The authors' systematic approach to this problem is valuable.

Technical Points and Interpretation

The authors model a large-diameter thin-walled cylinder, with the internal surface being overlaid with stainless steel. The key parameters they investigate are the welding sequence, the welding speed, the end clearance (the distance from the cylinder end to the start of the weld), and the number of overlay layers.

The following table summarizes the key findings.

Parameter Range Investigated Effect on End Deformation Practical Recommendation
Welding sequence From edge inward vs. from center outward Welding from the edge inward produces the smallest deformation. Always start the weld at the cylinder end and progress toward the center.
Welding speed 10–25 mm/s Higher speed generally reduces deformation, but there is an anomaly at 21.360 mm/s. Use the highest speed that allows a sound weld, but avoid the anomalous speed range.
End clearance 0–83.334 mm Increasing the clearance reduces the deformation. Provide a sufficient clearance to allow the end to move freely.
Number of layers 1–4 Deformation increases with the number of layers. Minimize the number of layers by using a thicker single layer if possible.

The most interesting finding is the anomalous behavior at a welding speed of 21.360 mm/s. The authors report that the deformation is abnormally high at this speed, and that normal welding cannot be achieved. This is likely due to a resonance or instability in the welding process at this specific speed, which causes an irregular heat input and a non-uniform temperature distribution. This is a valuable observation, as it highlights the importance of avoiding certain process parameters that can lead to unexpected behavior.

The authors also show that the welding sequence has a significant effect. When each layer is welded from the edge of the cylinder inward, the deformation is minimized. This is because the weld starts at a point where the cylinder is free to move, and the progressive welding creates a more uniform stress distribution.

Practical Insights and Lessons Learned

This paper is directly relevant to the fabrication of large-diameter pipe with an internal CRA overlay. In my experience, the distortion of the pipe ends is a major problem, especially for thin-walled pipe with a diameter greater than 500 mm. The ends must be within a tight tolerance to allow for the subsequent girth welding and assembly.

The authors' finding that the welding sequence should start at the edge and progress inward is consistent with my own experience. However, I would add that the welding sequence must also be symmetrical around the circumference. If the welder starts at one point and progresses around the cylinder, the cylinder will bow. The correct approach is to use a back-step sequence, where the weld is made in segments, and the segments are filled in a balanced manner.

The effect of the end clearance is also important. In practice, this means that the overlay should not be applied all the way to the end of the cylinder. A certain margin must be left, and the overlay should be stopped at a distance from the end. This margin is often specified in the engineering drawing, but it is sometimes ignored by the production team to save time. The authors' results show that this is a false economy.

The anomalous behavior at a welding speed of 21.360 mm/s is a reminder that the welding process is not always predictable. This could be due to a specific interaction between the welding parameters and the geometry of the cylinder, such as a natural frequency of the cylinder being excited by the welding arc. In practice, this means that the welding speed must be carefully selected and that a trial run is always advisable.

I would also note that the simulation does not account for the clamping conditions. In practice, the cylinder is often held in a fixture or on rollers, which constrains the deformation. The simulation results should be interpreted with this in mind. The actual deformation will be less than the simulated value if the cylinder is clamped.

Conclusion and Reflection

This paper is a valuable contribution to the practical literature on the overlay welding of thin-walled cylinders. The authors have systematically investigated the key process parameters and have provided clear recommendations for minimizing end deformation. The finding that the welding sequence should start at the edge and progress inward is particularly useful, as is the observation that a certain end clearance is necessary. The anomalous behavior at a welding speed of 21.360 mm/s is a cautionary note that highlights the need for careful process development and trial runs. For engineers involved in the fabrication of large-diameter pipe, vessels, or separators, this paper provides a solid basis for developing a welding procedure that minimizes distortion and facilitates assembly. I would recommend that the simulation results be validated with a limited set of welding trials on actual production components, as the boundary conditions in a real production environment are often different from those assumed in the model. The practical recommendations from this paper, when combined with proper fixturing and a disciplined welding sequence, can significantly reduce the rework and schedule delays that are all too common in this type of fabrication.