Numerical Simulation of Dissimilar Metal Surfacing Welding Between Low-Alloy Steel and Stainless Steel
Literature Overview and Core Intent
This paper from Chongqing University addresses a topic that is central to the nuclear power industry but has direct relevance to our field: the residual stress distribution in dissimilar metal welds. The authors have developed a thermal-elastic-plastic finite element method using ABAQUS to simulate the temperature field and stress field during surfacing welding. The key finding is that the moving heat source model based on a Gaussian surface heat source fits the actual molten pool shape well, and that there is a clear residual stress discontinuity at the interface between the dissimilar metals.
The reason this matters to me is that dissimilar metal welds are everywhere in our industry. We weld carbon steel to stainless steel, low-alloy steel to austenitic stainless steel, and we do it in piping systems, pressure vessels, and equipment components. The residual stresses that develop at these interfaces are not just a theoretical concern; they are a direct driver of stress corrosion cracking, fatigue failure, and premature component failure.
Technical Interpretation and Process Points
The development of a thermal-elastic-plastic finite element model is not a simple task. The authors had to account for temperature-dependent material properties, the moving heat source, and the complex geometry of the weld. The fact that they used a Gaussian surface heat source model is significant because it more accurately represents the heat input distribution from a welding arc than a simple uniform heat source.
The most important technical finding is the residual stress discontinuity at the interface. This is caused by the difference in thermal expansion coefficients and mechanical properties between the low-alloy steel and the stainless steel. When the weld cools, the two materials contract at different rates, and the interface becomes a site of high stress concentration.
The paper also demonstrates that in multi-pass welding, the subsequent passes have an annealing effect on the previously deposited passes. This is a practical insight. When you are laying down multiple layers of surfacing, the heat from the later passes can partially relieve the residual stresses in the earlier passes. However, this annealing effect is not uniform, and the final pass always leaves a high residual stress state.
The numerical model allows the authors to predict the residual stress distribution before actually performing the weld. This is invaluable for process design. You can test different welding sequences, heat input levels, and preheat temperatures without consuming expensive test coupons.
Integration with Engineering Practice
In the pipe industry, we encounter dissimilar metal welds in several critical applications. The most common is the transition joint between a carbon steel pipe and a stainless steel pipe, often used in process piping where the fluid changes from a corrosive to a non-corrosive environment. We also see this in cladding operations, where a low-alloy steel pipe is surfaced with a corrosion-resistant alloy.
The residual stress issue is particularly acute in these applications because the service environment often contains chlorides or other stress corrosion cracking agents. A high residual stress at the interface, combined with a susceptible microstructure, can lead to rapid failure. I have seen cases where a dissimilar metal weld failed within months of installation due to stress corrosion cracking that initiated at the interface.
The annealing effect of subsequent passes is something that we use in practice, but often without fully understanding the mechanics. When we are cladding a pipe with multiple layers, we know that the final layer will have the highest residual stress, but we do not always appreciate that the intermediate layers have been partially stress-relieved by the subsequent passes. This has implications for post-weld heat treatment. If you are relying on the annealing effect to reduce residual stress, you need to ensure that the final pass is not the one that creates the highest stress.
The numerical model also provides a way to optimize the welding sequence. In a multi-pass weld, the order in which the passes are deposited can have a significant effect on the final residual stress distribution. The model allows you to test different sequences and select the one that minimizes the peak residual stress.
Key Questions and Reflections
One question I would raise is the validation of the model against actual experimental measurements. The paper states that the Gaussian heat source model fits the molten pool shape well, but I would like to see a comparison of the predicted residual stresses with those measured by X-ray diffraction or neutron diffraction. Numerical models are only as good as their validation.
Another practical concern is the effect of welding consumables. The paper appears to model the weld metal as having properties similar to the stainless steel, but in practice, the filler metal composition can significantly affect the residual stress distribution. A nickel-based filler, for example, has a different thermal expansion coefficient than an austenitic stainless steel filler, and this can shift the stress distribution.
The model also assumes a perfect bond at the interface. In reality, there is always a transition zone where the composition changes from the low-alloy steel to the stainless steel. This transition zone can have properties that are significantly different from either parent material, and it can be a site of localized stress concentration.
Study Insights and Implications
The key insight from this paper is that residual stress in dissimilar metal welds is not just a function of the welding process; it is a function of the material properties and the geometry of the joint. This means that you cannot simply transfer a welding procedure from one application to another without considering the specific materials involved.
For the pipe industry, this has direct implications for the design of transition joints. We need to consider not just the mechanical strength of the joint, but also the residual stress state. A joint that is mechanically sound but has high residual stress may fail prematurely in service.
The numerical model also provides a tool for optimizing the welding procedure. By simulating different welding sequences, heat inputs, and preheat temperatures, we can identify the combination that minimizes residual stress before we ever strike an arc. This is a significant cost and time saving.
Reference Value and Outlook
This paper is an excellent reference for anyone involved in dissimilar metal welding. The methodology is sound, and the findings are directly applicable to industrial practice. The development of a reliable numerical model for residual stress prediction is a significant step forward.
In the future, I expect to see more widespread use of these models in the pipe industry. As computational power increases and models become more sophisticated, they will become a standard tool for welding procedure development. The ability to predict residual stress before welding is a game-changer for critical applications.
The paper also highlights the importance of understanding the interaction between material properties and welding process. This is a lesson that we need to internalize. We cannot treat welding as a simple heat input process; we need to understand the metallurgical and mechanical consequences of every weld we make.
Zhuojin Pipe Fitting Co., Ltd