Dynamic Simulation of Dissimilar Material Surfacing Temperature Field Based on ANSYS
Why Simulation Matters in the Shop
When you weld two different materials together, the temperature field is the master variable. It controls the residual stress, the distortion, the microstructure of the HAZ, and the dilution of the overlay. If you can predict the temperature field, you can predict the outcome. If you cannot, you are welding blind. The paper by Dong Xiaoqiang and Duan Hongyan at Shenyang University of Technology addresses the specific problem of dissimilar material surfacing, which is more complex than same-material welding because the thermal properties—thermal conductivity, specific heat, and density—are different on each side of the interface. The temperature field is asymmetric, and the residual stress distribution is more complex.
The Simulation Approach
The authors used ANSYS with the APDL programming language to apply a moving heat source and simulate the transient temperature field. The key technical elements are:
| Element | Description |
|---|---|
| Finite element model | 3D model with adaptive mesh refinement at the weld zone |
| Heat source model | Moving Gaussian heat source applied with APDL |
| Material properties | Temperature-dependent properties for both materials |
| Boundary conditions | Convection and radiation from the surface |
| Solution method | Transient analysis with automatic time stepping |
The adaptive mesh refinement is the most important technical contribution. The weld pool and the HAZ have steep temperature gradients, so the mesh must be fine in those regions. The authors used an automatic mesh generation technique that refines the mesh at the weld metal and coarsens it away from the weld. This reduces the computational time without sacrificing accuracy. The APDL program applies the heat source as a moving load, which is the standard approach for welding simulation.
The validation is the critical part. The authors compared the simulated temperature field with measured temperatures from a thermocouple test, and the results were close. This is what makes the simulation a practical engineering tool rather than a theoretical exercise. If the simulation matches the measurement, you can trust it to predict the temperature field for other process conditions.
Practical Application in the Shop
The practical value of this simulation is in process development. When you are developing a dissimilar material weld, you have to choose the welding parameters—current, voltage, travel speed, preheat, and interpass temperature. Each combination produces a different temperature field, which produces a different residual stress and microstructure. The traditional approach is to weld test coupons and measure the results. This is expensive and slow. The simulation allows you to screen a wide range of parameters without welding a single coupon. You can identify the process window that gives the lowest residual stress and the most uniform microstructure, and then verify it with a limited number of test coupons.
The simulation is also valuable for predicting distortion. Dissimilar material welds distort more than same-material welds because the thermal expansion coefficients are different. The simulation can predict the distortion, and you can use that to plan the fixturing and the machining allowance. I have seen parts scrapped because the distortion was not accounted for. The simulation would have caught it.
What the Simulation Does Not Tell You
This is where I want to be candid. The simulation gives you the temperature field, but it does not give you the residual stress, the microstructure, or the mechanical properties. Those require additional models and additional validation. The temperature field is the input to the stress and microstructure models, but the stress model requires a constitutive model that accounts for the temperature-dependent yield strength, the thermal expansion, and the phase transformations. The microstructure model requires a CCT diagram for the specific materials. These are not trivial to obtain.
The simulation also does not account for the weld pool fluid dynamics, which affect the shape of the weld pool and the dilution. The Gaussian heat source is a simplification. A more accurate model would use a double-ellipsoid heat source or a computational fluid dynamics model. The authors used a Gaussian source, which is adequate for a first-order approximation, but it does not capture the full complexity of the weld pool.
Lessons from the Field
I have used welding simulation in my own work, and I have learned that it is a tool, not a replacement for experience. The simulation is only as good as the input data. If the material properties are wrong, the simulation is wrong. If the heat source model does not match the actual process, the temperature field is wrong. The validation step is essential. You have to measure the temperature in the actual weld and compare it with the simulation. If they match, you can use the simulation with confidence. If they do not, you have to adjust the model.
The other lesson is that simulation is most valuable when it is used to screen parameters, not to replace testing. You simulate a hundred combinations, you narrow it down to five, and you test those five. This saves time and money without sacrificing confidence. The paper demonstrates this approach, and it is a good example of how simulation can be integrated into the process development workflow.
Closing Thoughts
This paper is a solid contribution to the practical literature on welding simulation. The adaptive mesh refinement and the moving heat source are standard techniques, but the application to dissimilar materials is a valuable extension. The validation against measured temperatures is what makes it a practical tool. If you are developing a dissimilar material weld, this simulation approach can save you time and money. The key is to validate the model against actual measurements before you trust it. Simulation is a guide, not a gospel.
Zhuojin Pipe Fitting Co., Ltd