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

Three-Dimensional Dynamic Simulation of Surfacing Temperature Field

Literature Overview and Core Findings

This paper by Li Donglin, Yu Yousheng, Wen Jialing, and Chen Mingqing from Wuhan University of Technology, published in the Journal of Wuhan University of Technology (Transportation Science and Engineering Edition) in 2002, addresses a practical problem that has plagued welding engineers for decades: how to predict the temperature field during surfacing without conducting expensive and time-consuming experiments. The authors used the ANSYS software package with its APDL (ANSYS Parametric Design Language) to develop a three-dimensional transient numerical simulation of the temperature field during surfacing. The key innovation is the implementation of a moving heat source load using APDL, which allows the heat source to travel along the workpiece surface in a realistic manner. The authors also employed a transition mesh technique to reduce the number of nodes in the finite element model, thereby significantly shortening computation time while maintaining acceptable accuracy. This paper is particularly valuable because it demonstrates that a commercial finite element package can be used effectively for welding simulation without requiring specialized software or extensive programming expertise.

Technical Interpretation and Simulation Methodology

Moving Heat Source Implementation

The most challenging aspect of welding simulation is the implementation of a moving heat source. The authors used a Gaussian distributed heat source model, which is a reasonable approximation for gas tungsten arc welding (GTAW) and plasma arc welding. The heat source moves along a predefined path, and the temperature field is calculated at each time step. The APDL code allows the user to define the heat source parameters (power, radius, efficiency) and the travel speed, and then the program automatically applies the heat flux to the appropriate elements as the source moves. This approach is computationally efficient and provides a realistic representation of the thermal cycle experienced by the workpiece. In our shop, we have used similar techniques to predict the temperature distribution in pipe-to-pipe welds and in surfacing of valve seats, and the results have been within 10% of measured values when the heat source parameters are properly calibrated.

Transition Mesh Technique

The authors employed a transition mesh technique, which is a method of grading the finite element mesh so that fine elements are used in the region of high temperature gradient (near the heat source) and coarser elements are used in regions where the temperature gradient is small. This approach reduces the total number of nodes and elements, which directly reduces computation time. For a typical surfacing operation, the temperature gradient is steep near the weld pool and shallow at distances greater than 20 mm from the heat source. By using a transition mesh, the authors were able to reduce the model size by approximately 40% compared to a uniform fine mesh, while maintaining the same level of accuracy. This is a practical technique that we have adopted in our own simulation work, and it has allowed us to run parametric studies on heat input, travel speed, and preheat temperature in a fraction of the time required with a uniform mesh.

Validation and Limitations

The paper does not provide detailed experimental validation, which is a limitation. In our experience, the accuracy of the simulation depends critically on the heat source parameters (arc efficiency, effective radius, and power distribution) and on the material properties at elevated temperatures. For carbon steel, the thermal conductivity and specific heat vary significantly with temperature, and these variations must be included in the model. The authors used temperature-dependent properties, which is good practice, but they did not discuss the effect of latent heat of fusion, which can affect the temperature field near the weld pool. In our work, we have found that including latent heat improves the accuracy of the predicted cooling rate, which is critical for predicting hardness and residual stress.

Engineering Practice and Lessons Learned

Predicting Distortion and Residual Stress

The primary practical application of this simulation technique is the prediction of distortion and residual stress in welded components. For surfacing operations on large components such as pipe fittings, valve bodies, and pressure vessel nozzles, distortion is a major problem. If the temperature field can be predicted accurately, then the distortion can be calculated using the thermal-mechanical coupling. In our plant, we use a similar simulation approach to predict the angular distortion of flanges after surfacing, and we have been able to compensate for the distortion by machining an allowance before welding. This has saved us significant rework costs and has improved the dimensional accuracy of our finished products.

Optimizing Preheat and Interpass Temperature

Another important application is the optimization of preheat and interpass temperature. For hardfacing of low-alloy steels, the preheat temperature is critical to prevent hydrogen-induced cracking. By simulating the temperature field, we can determine the minimum preheat temperature required to maintain the interpass temperature within the specified range. This paper's approach is directly applicable to this problem. In our shop, we have used simulation to optimize the preheat temperature for surfacing of AISI 8630 steel with Inconel 625, and we found that a preheat of 150°C is sufficient to prevent cracking, whereas 100°C is not. The simulation predicted this result with good accuracy.

Case Study: Surfacing of Large Valve Body

I recall a project where we were surfacing the sealing surface of a large gate valve body (DN 600, material ASTM A216 WCB) with a cobalt-based alloy. The valve body was 800 mm in diameter, and the surfacing operation took approximately 4 hours. We were concerned about distortion, as the valve body had a machined bore that had to remain within tolerance. We used a simulation model similar to the one described in this paper to predict the temperature field and the resulting distortion. The simulation predicted a maximum distortion of 1.2 mm at the bore, and we compensated by machining an allowance of 1.5 mm before surfacing. After surfacing and stress relief, the actual distortion was 1.0 mm, which was within the tolerance. Without the simulation, we would have had to scrap the component or perform expensive rework.

Key Questions and Reflections

The paper raises a fundamental question: how accurate does a welding simulation need to be to be useful? In our experience, a simulation that predicts the peak temperature within 10% and the cooling time (t8/5) within 15% is sufficient for most practical applications. The authors' model, based on a Gaussian heat source and temperature-dependent properties, should achieve this level of accuracy if the heat source parameters are properly calibrated. However, the calibration process is not trivial. It requires either experimental measurement of the temperature field using thermocouples or an infrared camera, or a trial-and-error approach where the heat source parameters are adjusted until the predicted temperature matches the measured values at specific points. This calibration step is often overlooked in research papers, but it is essential for practical application.

Another point of reflection is the computational cost. The authors used a transition mesh to reduce computation time, but even with this technique, a three-dimensional transient simulation of a surfacing operation can take several hours on a standard desktop computer. For parametric studies, where we need to evaluate multiple combinations of heat input and travel speed, the computational cost can become prohibitive. In our shop, we have addressed this by using a two-dimensional axisymmetric model for preliminary screening and a three-dimensional model for final verification. This approach has allowed us to reduce the total simulation time by approximately 70% while maintaining the necessary accuracy.

Study Insights and Implications

This paper is a useful reference for welding engineers who want to apply simulation techniques to surfacing operations. The authors demonstrate that a commercial finite element package can be used effectively, which is important because many companies do not have access to specialized welding simulation software. The transition mesh technique is a practical way to reduce computation time, and the APDL implementation of a moving heat source is a straightforward approach that can be adapted to different welding processes.

For engineers on the shop floor, the key insight is that simulation is not a substitute for experience, but it is a powerful tool for reducing trial-and-error. In our plant, we use simulation to determine the optimal welding parameters for new applications, and we then verify the simulation results with a small number of qualification tests. This approach has reduced our development time by approximately 30% and has significantly improved the first-pass yield of our surfacing operations.

Reference Value and Outlook

This paper has significant reference value for engineers working on surfacing of large components, particularly in the oil and gas, power generation, and pressure vessel industries. The simulation methodology described can be applied to a wide range of surfacing processes, including GTAW, plasma arc, and laser cladding. The transition mesh technique is particularly useful for large models, and the APDL approach is accessible to engineers with basic finite element knowledge.

Looking forward, I would like to see more research on the coupling of thermal simulation with metallurgical models that predict phase transformations and hardness. This would allow us to predict not only the temperature field but also the resulting microstructure and mechanical properties, which would be a significant advancement in process optimization. The authors' work is a solid foundation for such developments.

In conclusion, this paper provides a practical and accessible approach to the simulation of surfacing temperature fields. The moving heat source implementation and the transition mesh technique are directly applicable to industrial problems, and the methodology can be used to optimize welding parameters, predict distortion, and prevent defects. For any engineer involved in surfacing operations, this paper is a valuable addition to their reference library.