Influence of welding sequences on the accumulation evolution of angular distortion in thick-plate T-joint GMAW: Numerical simulation and modeling perspectives

In the heavy industry that builds ships, bridges, and massive pressure vessels, engineers rely on thick steel plates to hold immense weight and withstand extreme forces. To join these plates, workers use a process called welding, where intense heat melts the metal to fuse it together. However, this heat creates a hidden problem: as the molten metal cools and shrinks, it pulls the surrounding steel out of shape. In a specific type of joint known as a T-joint, where one plate stands vertically against another like the letter T, this shrinking causes the vertical plate to bend or tilt, a defect known as angular distortion. If this bending is too severe, the parts will not fit together correctly, compromising the safety and strength of the final structure. For decades, manufacturers have tried to manage this by changing the order in which they apply the welds, but predicting exactly how the metal will behave in thick sections has remained a difficult puzzle.

A team of researchers from Nanchang Jiaotong Institute and Nanchang University in China recently tackled this puzzle by studying a forty-millimeter-thick steel T-joint, a size common in heavy machinery. They focused on a specific design called a K-groove, which is a V-shaped cut made on both sides of the joint to ensure the weld penetrates deep enough to be strong. The team wanted to understand how the sequence of welding passes—the individual beads of molten metal laid down one after another—changes the way the distortion builds up over time. To do this, they first built physical models of the joints and welded them using three different sequences. They then created a detailed computer simulation that mimicked the heat and stress of the real welding process, allowing them to test six additional sequences that would have been too time-consuming to build by hand. Their goal was not just to find the best method, but to map out the entire life cycle of the distortion, watching how it grew, shifted, and settled as the welds were added.

The researchers discovered that the distortion does not happen all at once; instead, it evolves through two distinct phases. The first phase, which covers the first forty percent of the welding passes, is a period of intense adjustment. During this time, the vertical plate swings back and forth significantly as each new bead of hot metal is added and then cools. The second phase, comprising the remaining forty-eight percent of the passes, is a period of gradual convergence. Here, the metal becomes stiffer, and the swinging motion slows down, eventually settling into a final, stable angle. The team found that the most critical time to control the shape is during that initial, chaotic phase. They calculated that the ability of each new weld to correct the angle is roughly thirteen times stronger in the early adjustment phase than in the later, settling phase. This means that if a manufacturer wants to keep the joint straight, they must get the sequence right from the very beginning, rather than trying to fix it later.

When the researchers compared different ways of ordering the welds, they found that the most effective strategy involved a specific pattern of crossing over. Instead of welding one entire side of the joint before moving to the other, or simply alternating side-by-side, the best results came from a diagonally intersecting approach. In this method, the welder starts on the web side of one face, then moves to the opposite side to weld the corresponding layer, creating a diagonal path through the joint. This technique kept the final distortion to a minimum, reducing it by nearly forty percent compared to other common methods. In contrast, the approach of welding one entire side completely before touching the other side produced the worst results, causing the joint to tilt significantly. The study also revealed that while alternating every single pass offered the best control over the shape, it required more time and effort to switch sides constantly. A slightly less precise but more efficient method was to alternate by layers, completing a full layer on one side before switching, which still offered good control without the constant interruption.

The team also looked at the invisible forces left behind in the metal, known as residual stress. They found that the welding sequence that minimized the bending also resulted in the smallest areas of high tension within the steel. This is important because high tension can make metal more prone to cracking over time. The simulations showed that the best sequences created a more balanced distribution of stress, whereas the poor sequences left large pockets of intense pressure. By combining their physical experiments with the computer models, the researchers were able to create mathematical descriptions of how the distortion grows and how much each layer of weld can correct it. These descriptions act like a guidebook, showing that the process follows a predictable curve that can be modeled to anticipate the final shape.

Ultimately, this work provides a clear roadmap for engineers building thick steel structures. It confirms that the order in which the welds are applied is just as important as the heat used to make them. By understanding that the metal goes through a wild, swinging phase before it settles, and by using a diagonal crossing pattern to manage that swing, manufacturers can produce stronger, straighter joints with less waste and fewer repairs. The study suggests that for the best results, the welding plan should focus heavily on the first forty percent of the work, using a sequence that constantly balances the heat on both sides of the joint. This approach turns a complex, unpredictable problem into a manageable process, ensuring that the massive steel components holding up our world remain true to their design.