Dynamic Process of Welding Angular Distortion - Numerical Simulation and Experimental Study
Literature Overview and Core Message
This paper by Liu Chuan, Wang Rui and Zhang Jianxun from Xi'an Jiaotong University addresses a problem that has cost me more schedule days than almost anything else in fabrication: angular distortion in welded assemblies. The authors used finite element simulation to study the dynamic process of angular distortion during CO2 gas shielded arc welding, and they compared three modelling approaches: a two-dimensional contact constraint model, a two-dimensional displacement constraint model, and a three-dimensional displacement constraint model. The crucial innovation is that they modelled the fixture and the workpiece together in one finite element model, so that the contact state between the clamp and the part could be represented realistically rather than assumed. The results are clear and, to an experienced fabricator, very familiar.
Interpretation of Technical Points
The most important finding is that the two-dimensional models, whether contact constraint or displacement constraint, cannot capture the dynamic evolution of angular distortion during the welding process, but they can predict the final residual angular distortion with acceptable accuracy when compared with experimental results. The three-dimensional displacement constraint model can represent the dynamic distortion behaviour, but at a significant computational cost. The contact constraint model, which includes the fixture in the model, is the only approach that can calculate the friction force and contact pressure between the workpiece and the clamp. This matters because the clamping force and the friction between the part and the fixture are precisely what control how much distortion is allowed to develop during welding, and how much is locked in as residual distortion after the clamps are released.
Connection with Manufacturing Practice
In my own pipe fabrication shops, angular distortion is a daily battle. When we weld a flange to a pipe end, or a gusset plate to a large-diameter pipe, or a stiffener ring to a fitting, the weld shrinkage pulls the material out of square, and if we do not control it, the flange face will be out of perpendicularity and the whole assembly will be rejected at final inspection. The paper's finding that a two-dimensional model can predict the residual distortion is useful for design of experiments, but the practical value lies in the contact constraint model, because it tells us how much the clamping arrangement actually resists the distortion. I have seen shops use massive clamps and still get distortion, because the clamp was applied in a way that allowed the part to slide or rotate slightly under the weld shrinkage force. The friction between the fixture and the workpiece is the real resistance, and if the clamp surface is greasy or worn, the friction drops and the distortion increases.
Key Questions and Reflections
The paper does not address the effect of welding sequence, which in multi-pass welding is often more important than the clamping arrangement. In a thick-wall pipe joint, the angular distortion is built up pass by pass, and the order in which the passes are deposited can be used to balance the shrinkage and minimise the final distortion. I have used back-step welding and balanced welding sequences on large-diameter pipe joints with great success, and I would like to see a numerical model that incorporates the welding sequence as a variable. The paper also uses a relatively simple CO2 gas shielded welding process, and the results may not transfer directly to submerged arc welding or to multi-process welding, where the heat input and the deposition rate are very different. The cooling time and the shrinkage force are both functions of the heat input, so the distortion behaviour will differ.
Study Insights and Implications
For my own fabrication management, the practical lesson is that we should not rely on the clamping arrangement alone to control angular distortion. We need a combination of proper fixture design, adequate clamping force with clean and dry contact surfaces, and a welding sequence that balances the shrinkage. The numerical simulation approach described in this paper is a valuable tool for optimising the fixture design before we build it, and for predicting the residual distortion so that we can plan the machining allowance accordingly. I have used similar simulations to decide whether a particular flange-to-pipe weld needs a distortion control fixture or whether a simple tack weld and a skilled welder will suffice, and the simulation saved us from building an expensive fixture that was not needed. In summary, this paper is a good example of how numerical simulation can be used to solve a practical fabrication problem, and it should be read by anyone who is responsible for dimensional control in welded assemblies.
Zhuojin Pipe Fitting Co., Ltd