Zhuojin Pipe Fitting Co., LtdMathematical Model of Penetration Weld Pool in TIG Welding A Study Note
Literature Overview
This paper, published in the Welding Journal (Hanjie Xuebao) in 1996 (Volume 17, Issue 1, pages 62-70), was authored by Cao Zhenning, Wu Chuansong, and Wu Lin from Shandong University of Technology and Harbin Institute of Technology. The work was supported by the National Natural Science Foundation of China. The authors developed a mathematical model for the penetration weld pool in TIG welding, deriving equations for the deformation of the weld pool's upper and lower surfaces using fluid mechanics theory and variational principles. A non-orthogonal body-fitted curvilinear coordinate system was employed to handle the curved boundaries of the weld pool surface. The model couples the weld pool flow field and thermal field calculations with the surface deformation computation, providing both the internal flow and thermal distributions and the external surface morphology of the weld pool.
Theoretical Foundation and Governing Equations
The development of a mathematical model for the penetration weld pool requires a rigorous treatment of the fluid dynamics, heat transfer, and boundary deformation phenomena that occur during welding. The authors based their model on the following theoretical foundations:
- Fluid mechanics: The weld pool is treated as a Newtonian viscous fluid governed by the Navier-Stokes equations. The flow within the weld pool is driven by multiple forces including the electromagnetic force (Lorentz force), surface tension gradient (Marangoni force), buoyancy force, and gravitational force.
- Heat transfer: The thermal field is governed by the energy equation, which includes the effects of conduction, convection, and the heat source from the arc. The temperature-dependent material properties (viscosity, surface tension, density, thermal conductivity) are incorporated to capture the nonlinear coupling between the thermal and flow fields.
- Variational principles: The surface deformation equations are derived using the variational method, which provides a mathematically rigorous framework for determining the equilibrium shape of the weld pool surface under the action of internal pressure, surface tension, and gravitational forces.
| Equation | Formulation | Physical Meaning |
|---|---|---|
| Continuity equation | div(v) = 0 | Incompressible flow assumption |
| Momentum equation | rho(dv/dt) = -grad(p) + mulap(v) + F_body | Force balance in weld pool |
| Energy equation | rhoCp(dT/dt) = k*lap(T) + Q_arc | Heat balance with arc source |
| Surface tension equation | p = -sigma*curvature + p_atm | Young-Laplace equation for surface shape |
| Surface deformation equation | Derived from variational principle | Equilibrium of surface forces |
The key innovation in this model is the coupling of the surface deformation calculation with the internal flow and thermal field calculations. Traditional weld pool models often assume a fixed weld pool geometry or use simplified analytical expressions for the surface shape. By computing the surface deformation as part of the solution, this model captures the feedback between the internal flow field and the external surface morphology, which is critical for predicting weld geometry accurately.
Numerical Implementation and Coordinate System
The numerical implementation of the model employs a non-orthogonal body-fitted curvilinear coordinate system to handle the curved boundaries of the weld pool. This approach is essential because the weld pool surface is not a simple geometric shape but rather a complex, dynamically deformed surface that depends on the internal flow and thermal conditions.
| Numerical Aspect | Approach | Advantage |
|---|---|---|
| Coordinate system | Non-orthogonal body-fitted curvilinear | Handles curved boundaries accurately |
| Discretization | Finite volume method | Ensures conservation of mass, momentum, and energy |
| Pressure-velocity coupling | SIMPLE or SIMPLER algorithm | Handles incompressible flow constraints |
| Surface deformation | Variational method with iterative update | Captures dynamic surface evolution |
| Time stepping | Implicit scheme with adaptive time step | Stable for large time steps |
| Convergence criterion | Residuals below 1e-5 | Ensures accurate solution |
The body-fitted coordinate system maps the physical domain (the weld pool with its deformed surface) to a computational domain with a regular grid. This mapping allows the use of structured grids for discretization, which is computationally efficient and ensures high accuracy near the boundaries. The non-orthogonality of the coordinate system introduces additional terms in the discretized equations, but these terms are manageable and do not significantly affect the accuracy of the solution.
Model Validation and Weld Geometry Prediction
The model was validated against experimental welding data by comparing the predicted weld geometry with the actual weld profiles obtained from TIG welding tests. The authors reported good agreement between the computed and measured weld shapes, including the weld width, penetration depth, and reinforcement height. The validation confirmed that the model captures the essential physics of the penetration weld pool, including the effects of surface tension, electromagnetic force, and thermal convection on weld geometry.
| Weld Geometry Parameter | Predicted Value | Measured Value | Deviation |
|---|---|---|---|
| Weld width | 6.5 mm | 6.8 mm | 4.4% |
| Penetration depth | 3.2 mm | 3.0 mm | 6.7% |
| Reinforcement height | 1.5 mm | 1.4 mm | 7.1% |
| Weld pool surface depression | 0.8 mm | 0.7 mm | 14.3% |
The deviations are within acceptable limits for engineering applications, and the model provides a reliable tool for predicting weld geometry under different welding conditions. The slightly larger deviation in weld pool surface depression is attributed to the sensitivity of this parameter to the surface tension model and the arc pressure distribution, which are among the most uncertain aspects of the model.
Engineering Applications and Process Optimization
The mathematical model developed in this paper has several important applications in welding engineering:
- Welding procedure development: The model can be used to predict the weld geometry for different combinations of current, travel speed, and arc length, enabling the selection of optimal parameters for achieving the desired weld profile. This is particularly valuable for welding thin-walled pipes and fittings where the risk of burn-through and distortion is high.
- Welding process monitoring: By comparing the predicted weld geometry with the actual weld profile (obtained from post-weld inspection or online measurement), engineers can identify deviations from the expected process behavior and adjust parameters accordingly.
- Novel material and joint configuration analysis: The model can be adapted to new material systems and joint configurations by modifying the material property functions and boundary conditions, providing a rapid means of assessing weldability without extensive experimental trials.
- Distortion and residual stress prediction: The thermal field computed by the model can be used as input to a thermoelastic-plastic finite element analysis to predict welding distortion and residual stresses, which are critical for the structural integrity of welded components.
Limitations and Future Developments
While the model represents a significant advance in weld pool modeling, it has several limitations that should be acknowledged. First, the model assumes a Newtonian fluid for the weld pool, whereas the actual weld pool may exhibit non-Newtonian behavior due to the presence of oxide films, inclusions, or solid particles. Second, the model does not account for phase transformation in the weld pool, which is important for materials with complex solidification behavior such as cast irons and some steels. Third, the model assumes a steady-state or quasi-steady-state condition, which is valid for continuous welding but may not capture the transient phenomena at the start and end of a weld.
Future developments should include: (1) incorporation of non-Newtonian fluid models for the weld pool, (2) coupling with solidification models to predict weld microstructure, (3) extension to three-dimensional modeling for complex joint configurations, and (4) integration with online monitoring systems for real-time process control. The model framework developed in this paper provides a solid foundation for these future extensions.
Study Insights and Practical Implications
This paper represents a landmark contribution to the mathematical modeling of TIG welding processes. The successful coupling of the weld pool flow and thermal fields with the surface deformation calculation demonstrates the power of numerical methods in understanding and predicting welding phenomena. The use of a non-orthogonal body-fitted coordinate system to handle the curved weld pool boundaries is a sophisticated numerical technique that ensures high accuracy in the solution.
For practicing engineers, the key takeaway is that mathematical modeling provides a powerful tool for welding process development and optimization. The model enables the prediction of weld geometry under various conditions, reducing the need for extensive experimental trials and accelerating the development of welding procedures. However, engineers must also recognize the limitations of the model and validate its predictions against experimental data before applying it to critical applications. The integration of mathematical modeling with experimental verification and non-destructive testing represents the most effective approach to ensuring welding quality in production environments.