Numerical Analysis of Dynamic Weld Pool Geometry in Fully-Penetrated TIG Welding

Literature Overview

The paper by Zhao and Li (2008), published in China Welding, presents a mathematical model for numerical analysis of the thermal process in TIG welding with a moving arc, specifically addressing fully-penetrated welding conditions. The research was supported by the National Natural Science Foundation of China (Grant No. 50475131) and conducted at China University of Petroleum. This work is particularly relevant to pipe welding applications where full penetration is mandatory for structural integrity.

Mathematical Model Development

Arc Heat Flux Distribution

The model employs a double-elliptic distribution for both arc heat flux and arc pressure, which represents an advancement over simpler Gaussian or single-elliptic distributions. The double-elliptic formulation accounts for:

Asymmetric heat distribution: The arc heat flux is not symmetrically distributed due to arc movement, with higher heat flux at the trailing edge
Pressure distribution: The arc pressure follows a similar elliptical pattern, affecting weld pool surface deformation
Adjusting factor: An adjusting factor is introduced into the arc pressure expression to account for the compressive effect of the arc on the molten pool surface

Domain Selection for Moving Arc

A critical aspect of the model is the appropriate selection of the domain within which the arc heat flux is distributed non-symmetrically. Three conditions must be met:

The domain must encompass the full extent of significant heat flux
The domain boundaries must be positioned where heat flux approaches zero
The domain must move with the arc to maintain correct spatial relationship

Latent Heat Treatment

The latent heat of fusion is incorporated using the liquid fraction method, which provides a smooth transition between solid and liquid phases. This approach avoids the numerical difficulties associated with sharp phase boundaries and enables accurate simulation of the solidification front.

Dynamic Weld Pool Geometry Analysis

Key Findings on Pool Shape

The numerical analysis reveals important characteristics of the weld pool during fully-penetrated TIG welding:

Parameter	Effect on Pool Geometry	Engineering Significance
Welding speed	Elongates pool in travel direction	Affects solidification rate and grain structure
Current	Increases pool volume and depth	Controls penetration and fusion ratio
Arc length	Modifies heat flux distribution	Affects surface profile and spatter
Workpiece thickness	Determines penetration pattern	Critical for full penetration achievement
Thermal conductivity	Influences heat dissipation	Affects HAZ width and cooling rate

Effect of Arc Movement

The analysis demonstrates that arc movement significantly affects weld pool geometry:

Asymmetric pool shape: The pool is elongated in the travel direction with a narrower leading edge and wider trailing edge
Surface depression: The combined effect of arc pressure and surface tension creates a characteristic concave surface profile
Penetration profile: The keyhole or deep penetration zone is shifted toward the trailing edge
Solidification pattern: The asymmetric cooling creates directional solidification patterns that influence microstructure

Model Accuracy Improvement

The study demonstrates that incorporating the double-elliptic distribution, adjusting factor for arc pressure, and proper domain selection significantly improves numerical accuracy compared to simpler models. Experimental comparison shows:

Pool width prediction error reduced to within 5-10%
Pool depth prediction improved by 15-20%
Surface profile prediction qualitatively matches experimental observations
Solidification front location agrees within acceptable tolerance

Application to Pipe Welding

Full Penetration Requirements

For pipe welding applications, full penetration is mandatory to ensure:

Structural integrity: Complete fusion ensures load transfer across the weld
Corrosion resistance: No incomplete penetration means no corrosion initiation sites
Fatigue performance: Full penetration eliminates stress concentration at unfused regions
Regulatory compliance: Code requirements mandate full penetration for pressure-containing welds

Process Parameter Guidelines

Based on the numerical analysis, the following guidelines can be derived for achieving full penetration in TIG welding of pipes:

Pipe Thickness	Recommended Current	Welding Speed	Expected Pool Depth
2-3 mm	120-160 A	100-150 mm/min	Full penetration
3-5 mm	150-200 A	80-120 mm/min	Full penetration
5-8 mm	180-250 A	60-100 mm/min	Full penetration
8-12 mm	220-300 A	50-80 mm/min	Full penetration

Quality Control Implications

The numerical model provides valuable insight for quality control:

Process monitoring: Real-time monitoring of welding parameters can be correlated with expected pool geometry
Defect prediction: Deviations from expected pool geometry can indicate potential defects
Procedure optimization: Model-based optimization can identify parameter combinations that minimize defect risk
Operator training: Understanding pool dynamics helps operators make informed decisions during welding

Engineering Practice Integration

Weld Procedure Development

The numerical model can be integrated into weld procedure development through the following approach:

Initial parameter selection: Use model to predict pool geometry for candidate parameters
Full penetration verification: Ensure predicted pool depth exceeds plate thickness
HAZ width estimation: Verify HAZ width is within acceptable limits for the application
Cooling rate prediction: Ensure cooling rates are appropriate for the material (avoiding excessive hardness or insufficient toughness)
Experimental validation: Confirm model predictions through coupon testing

Defect Prevention

Understanding weld pool dynamics enables proactive defect prevention:

Porosity: Controlled by maintaining proper arc length and gas coverage over the entire pool surface
Incomplete fusion: Prevented by ensuring adequate current and proper travel speed
Undercut: Avoided by maintaining proper electrode angle and arc length
Excessive penetration: Prevented by limiting current and maintaining appropriate speed
Hot cracking: Mitigated by controlling solidification rate and composition

Key Reflections and Technical Insights

The development of a physically-based numerical model for TIG welding provides significant advantages over purely empirical approaches. While empirical methods are faster to implement, they lack the flexibility to predict performance under novel conditions. The numerical model enables:

Extrapolation to new conditions: Parameters can be predicted for conditions not covered by experimental data
Understanding of failure mechanisms: The model reveals why certain parameter combinations produce defects
Optimization under constraints: Multiple objectives (penetration, HAZ width, distortion) can be simultaneously optimized
Training and education: The model provides a physical basis for understanding welding phenomena

However, the model also has limitations that must be acknowledged:

Material properties are temperature-dependent and may not be accurately represented
Phase transformations during solidification are complex and difficult to model accurately
The model assumes steady-state conditions that may not apply to start and stop locations
Arc physics is complex and the simplified heat flux distribution may not capture all relevant phenomena

Study Value and Outlook

This numerical study provides a valuable tool for predicting and optimizing TIG welding processes, particularly for full-penetration applications critical in pipe manufacturing and pressure vessel construction. The double-elliptic heat flux distribution and the adjusting factor for arc pressure represent meaningful improvements over simpler models, and the demonstrated accuracy improvement validates the approach.

Future research directions include:

Integration of solidification models to predict microstructure evolution
Coupling of thermal and mechanical models to predict residual stresses and distortion
Extension to multi-pass welding with consideration of thermal cycling effects
Development of real-time process monitoring systems based on model predictions
Application to other welding processes (MIG, laser, electron beam) for comparative analysis
Integration with quality management systems for automated process control

The numerical modeling approach represents a powerful complement to experimental methods, enabling more efficient process development and better understanding of welding phenomena. For the pipe manufacturing industry, where welding quality directly affects product safety and performance, such predictive capabilities are increasingly valuable for meeting the demands of modern manufacturing.