Numerical Simulation of Wire-Feed TIG Welding on Stainless Steel Thin Plates

Literature Overview

The paper by Yang Wenfei and Liu Aiguo, published in the Journal of Shenyang Ligong University (Vol. 40, No. 6, 2021, pp. 71–76), presents a finite element simulation study of wire-feed TIG (WFTIG) welding on 2 mm thick 304 stainless steel plates using the Simufact Welding software. The study employs a simplified physical model with a double-ellipsoid heat source and investigates the effects of welding current and travel speed on temperature field, stress field, and weld distortion. This work is significant for understanding the thermal-mechanical behavior of thin stainless steel weldments, where excessive heat input can lead to unacceptable distortion and loss of dimensional accuracy.

Core Technical Context

Wire-feed TIG welding combines the advantages of conventional TIG (excellent weld quality, clean welds, no spatter) with the higher deposition rates of arc-wire processes. In WFTIG, a continuous filler wire is fed into the arc, allowing for single-pass welding of thin plates that would require multiple passes with conventional TIG. For 2 mm 304 stainless steel, this is particularly relevant in applications such as heat exchanger tubesheets, nuclear components, and aerospace structures where weld quality and distortion control are critical.

304 stainless steel presents unique welding challenges: its low thermal conductivity (approximately 16 W/m·K at room temperature) results in high heat concentration at the weld zone, its high thermal expansion coefficient (approximately 17.3 μm/m·K) amplifies thermal distortion, and its susceptibility to intergranular corrosion and sensitization requires careful control of the thermal cycle.

Simulation Methodology

The Simufact Welding software implements coupled thermo-mechanical finite element analysis, solving heat transfer and structural equations simultaneously. The double-ellipsoid heat source model represents the arc heat distribution with a keyhole component (for penetration) and a surface component (for spreading), providing a more realistic thermal profile than simpler models such as Gaussian or single-ellipsoid sources.

Key Simulation Results

Temperature Field Analysis

The simulation reveals that the peak temperature at the weld pool center increases by approximately 500 K when the welding current is raised from 120 A to 160 A. This significant temperature increase corresponds to a substantial expansion of the molten zone and a broader HAZ. For 304 stainless steel, peak temperatures exceeding 1400 °C in the weld pool and 800–1100 °C in the HAZ are typical, which can approach or exceed the sensitization temperature range (450–850 °C) if cooling rates are slow.

When travel speed increases from 12 cm/min to 16 cm/min, the peak temperature decreases by approximately 400 K. This reduction narrows the molten zone and reduces the extent of the HAZ, but may also increase the cooling rate, potentially leading to martensitic transformation in the weld metal if the composition is not properly controlled.

Stress Field and Distortion Analysis

The simulation results indicate that stress concentration is most pronounced at the weld start point. This is a critical finding because weld start and stop locations are common sites for defect initiation in engineering practice. The stress concentration at the start point arises from the abrupt introduction of thermal loading, which creates a localized stress gradient that exceeds the bulk weld stress.

A notable observation is that changing the welding current or travel speed has minimal effect on the location of the stress concentration zone. This implies that the start-point stress concentration is an inherent characteristic of the welding process geometry and cannot be eliminated by parameter adjustment alone. Process modifications such as back-step welding, tack welding, or the use of backing bars are more effective countermeasures.

The z-direction distortion at tracking point C shows a maximum increase of 0.06 mm when current increases from 120 A to 160 A, and a maximum decrease of 0.33 mm when speed increases from 12 cm/min to 16 cm/min. The much larger distortion reduction from speed increase compared to the distortion increase from current increase suggests that travel speed is the more effective parameter for controlling thin-plate distortion. This finding has direct implications for process optimization: for thin stainless steel plates, prioritizing higher travel speeds with correspondingly lower currents can significantly reduce angular and buckling distortion.

Engineering Practice Implications

For 2 mm 304 stainless steel, the simulation results support a process philosophy of moderate current with high travel speed. A current of 120–140 A with a travel speed of 14–16 cm/min would provide a good balance between penetration and distortion control. The heat input in this range would be approximately 3–5 kJ/mm, which is within the recommended range for thin stainless steel to minimize sensitization risk.

In practice, distortion control for thin stainless steel weldments requires a multi-pronged approach:

• Pre-weld fixtures and clamps to restrain the plate and limit angular distortion
• Back-step welding to reduce start-point stress concentration
• Interpass temperature control below 150 °C to limit HAZ grain growth
• Post-weld stress relief annealing at 400–500 °C for critical applications
• Use of backing bars to support the root and prevent sagging

The simulation's prediction that travel speed is the dominant distortion control parameter is consistent with engineering experience, where high-speed welding is routinely employed for thin stainless steel to minimize heat-affected distortion. However, the simulation does not account for the practical limitations of WFTIG at very high speeds, where arc stability and wire feed consistency may become challenging.

Key Questions and Reflections

The study's use of a simplified physical model raises questions about the accuracy of the simulation results for real-world conditions. The double-ellipsoid heat source, while more realistic than Gaussian sources, still does not fully capture the complex fluid dynamics of the weld pool, including convection currents, surface tension effects, and keyhole formation in wire-feed TIG. The absence of a moving mesh or remeshing strategy may also affect the accuracy of the temperature field near the weld pool boundary.

Additionally, the study does not address the metallurgical consequences of the simulated thermal cycles. For 304 stainless steel, the thermal cycle parameters directly influence the formation of chromium carbides at grain boundaries, which can lead to intergranular corrosion. A comprehensive study would correlate the simulated thermal cycles with predicted microstructural evolution and corrosion resistance.

Summary and Outlook

The numerical simulation of wire-feed TIG welding on 2 mm 304 stainless steel provides valuable quantitative insights into the thermal-mechanical behavior of thin-plate welds. The finding that travel speed is the dominant parameter for distortion control, with a 0.33 mm reduction achievable by increasing speed from 12 to 16 cm/min, offers clear guidance for process optimization. The persistent stress concentration at the weld start point highlights the need for process modifications beyond simple parameter adjustment. For engineers working with thin stainless steel components, these simulation results support a process strategy that emphasizes high travel speeds, moderate currents, and mechanical restraint to achieve both quality and dimensional accuracy.