CLADDING TECHNOLOGY SHANXI CO., LTDNumerical Simulation of Welding Temperature Fields and Deformation in Cladding Operations
Overview and Technical Context
The fabrication of bimetallic pressure vessels, particularly those involving dissimilar metal cladding of austenitic stainless steel or nickel-based alloys onto carbon and low-alloy steel substrates, presents formidable challenges in predicting and controlling welding-induced deformation and residual stress. Topic 360 addresses the application of finite element analysis (FEA) using commercial software platforms such as Sysweld and Simufact Welding to simulate the welding temperature field, cooling rates, residual stress distribution, and geometric deformation before any physical welding operation is performed. This approach enables the pre-planning of overlay sequence, cooling strategies, and backing arrangements, thereby significantly reducing trial-and-error costs during process development and new product introduction.
The motivation for numerical simulation in cladding operations is rooted in the fundamental incompatibility between the thermal expansion coefficients of the base metal and overlay material. For instance, 304 stainless steel has a linear thermal expansion coefficient of approximately 17.3 × 10⁻⁶ /°C, while typical carbon steel (SAE 1020) has a coefficient of about 11.7 × 10⁻⁶ /°C. This mismatch, combined with the directional heat input from multi-pass welding, creates complex stress states that are difficult to predict through empirical methods alone.
Core Simulation Methodology
The numerical simulation workflow for cladding operations typically follows a systematic sequence:
- Geometry modeling: Three-dimensional finite element models of the workpiece, including the base plate, backing plate, and planned overlay geometry, are created with appropriate mesh density concentrated near the weld zone.
- Material property definition: Temperature-dependent thermal conductivity, specific heat, yield strength, and thermal expansion coefficients for both base and overlay materials are input based on experimental data or established databases.
- Heat source modeling: The welding heat input is represented using analytical models such as the double-ellipsoidal Goldak model or the conical heat source model, calibrated against measured temperature profiles.
- Thermal analysis: Transient heat transfer analysis is performed to predict temperature field evolution, peak temperatures, cooling rates (particularly t₈/₅ cooling time), and thermal cycles experienced at critical locations.
- Mechanical analysis: Elastic-plastic analysis, often using the coupled thermo-mechanical approach or the deadloading method, predicts residual stress distributions and geometric deformations.
- Process optimization: Based on simulation results, the overlay sequence, cooling strategy, and restraint conditions are iteratively refined.
| Simulation Parameter | Typical Input Value | Sensitivity Level |
|---|---|---|
| Heat input (Q) | 15–40 kJ/mm | High |
| Travel speed (v) | 3–10 mm/s | High |
| Heat source eccentricity (a, b) | 0.4–0.8 | Medium |
| Thermal conductivity (k) | 15–40 W/(m·K) | Medium |
| Yield strength (σy) | 150–350 MPa (temperature-dependent) | High |
| Thermal expansion coefficient (α) | 11–18 × 10⁻⁶ /°C | High |
| Mesh size near weld | 0.5–2.0 mm | High |
| Time step | 0.01–0.1 s | Medium |
Process Optimization Through Simulation
One of the most valuable applications of numerical simulation in cladding operations is the optimization of the overlay welding sequence. For a typical multi-layer, multi-pass overlay on a large-diameter vessel shell, the sequence directly influences the final residual stress state and geometric distortion. Simulation allows engineers to evaluate multiple sequence alternatives virtually:
- Sequential pass-by-pass simulation: Each pass is modeled individually with appropriate cooling between passes, capturing the progressive stress buildup and relaxation.
- Symmetric welding sequences: Simulation confirms that symmetric welding patterns minimize angular distortion in butt-welded clad assemblies.
- Skip-welding strategies: For large surfaces, simulation helps determine optimal skip distances that balance distortion control with production efficiency.
- Backing arrangement analysis: The influence of copper backing plates, water-cooled backing, and ceramic backing blocks on cooling rates and distortion can be evaluated without physical trials.
The cooling rate prediction capability of simulation is particularly important for determining the metallurgical outcomes of the overlay layer. For austenitic stainless steel overlays, cooling rates below 50°C/s generally produce fully austenitic microstructures, while faster cooling rates may introduce delta ferrite that can improve crack resistance but reduce ductility. For nickel-based alloy overlays, cooling rate directly influences the precipitation state of strengthening phases, affecting both mechanical properties and corrosion resistance.
Integration with Engineering Practice
In a recent project involving the fabrication of a hydrogenation reactor with 316L stainless steel overlay on 16Mn low-alloy steel, numerical simulation was used to evaluate three alternative overlay strategies. The simulation revealed that a conventional sequential pass pattern would produce a maximum angular distortion of 3.2 mm per meter, exceeding the acceptable tolerance of 1.5 mm per meter. By implementing a simulated symmetric skip-welding pattern with intermediate cooling intervals, the predicted distortion was reduced to 0.8 mm per meter, well within specification. The physical welding operation, executed according to the simulation-optimized procedure, achieved a measured distortion of 0.9 mm per meter, validating the simulation accuracy to within 12%.
The economic benefits of simulation-driven process optimization are substantial. For a new product introduction involving a novel material combination, physical trial welding and full-scale testing can require 4–6 weeks and consume significant quantities of expensive alloy consumables. Simulation reduces this development cycle to approximately 2–3 weeks by identifying viable process windows before physical trials begin, with subsequent trial welding serving as confirmation rather than exploration.
Key Questions and Reflections
A critical question that emerges from this study is the accuracy boundary of numerical simulation. While modern FEA software can reproduce measured temperature fields and residual stresses with acceptable accuracy (typically within 10–15%), the predictive capability diminishes significantly for novel material combinations where temperature-dependent material properties are not well characterized. The accuracy of simulation is ultimately limited by the quality of input data, particularly the temperature-dependent yield strength and thermal expansion coefficients, which are often extrapolated from limited experimental data.
Another important consideration is the computational cost of full-scale simulation for large pressure vessels. A complete simulation of a 6-meter diameter reactor with multi-pass overlay on both internal and external surfaces can require hundreds of CPU hours. Practical approaches involve simulating representative sections with appropriate boundary conditions and scaling the results, or employing reduced-order modeling techniques that capture the essential physics while significantly reducing computational requirements.
The integration of simulation with real-time monitoring systems (as described in Topic 359) represents a promising direction for future development. By comparing measured thermal cycles with simulated predictions during actual welding operations, discrepancies can be identified and corrective actions taken in real time, creating a truly closed-loop process control system for cladding operations.