Detection of TIG Welding Molten Pool Geometric Parameters

Literature Overview

This 2015 study published in Hot Working Technology by Duan Ruixia, Xu Jianyu from Ningbo University's School of Information Science and Engineering, and Ke Yu from Ningbo Xinle Life Electric Appliance Co., Ltd. addresses the critical challenge of real-time detection and measurement of TIG welding molten pool geometric parameters. The research investigates optical sensing methods for monitoring weld pool width, depth, and profile during the welding process, with the aim of enabling closed-loop process control and quality assurance. The study bridges the gap between academic welding research and industrial application through collaboration between a university research group and a manufacturing enterprise.

Core Technical Content

Real-time molten pool monitoring is essential for ensuring weld quality because the pool geometry directly determines the final weld bead dimensions, penetration profile, and defect susceptibility. The study evaluates several optical sensing approaches for measuring key pool parameters:

Detection Method	Measurable Parameters	Accuracy	Speed	Cost
Visible light imaging	Pool width, surface profile	±0.5 mm	100–500 fps	Medium
Infrared thermography	Pool temperature distribution, approximate depth	±1.0 mm	50–200 fps	Medium-High
Laser displacement scanning	Pool depth, cross-sectional profile	±0.1 mm	10–50 Hz	High
Fiber optic sensing	Pool temperature, width	±0.3 mm	1000+ Hz	Low
Multi-sensor fusion	All parameters with cross-validation	±0.2 mm	50–200 fps	High

The researchers developed a multi-sensor fusion approach that combines visible light imaging for surface geometry with infrared sensing for thermal distribution, achieving comprehensive pool characterization with improved accuracy compared to single-sensor methods.

Methodology and Technical Approach

The study employs a systematic approach to molten pool parameter detection:

Sensor configuration: A visible light camera mounted at a 45-degree angle captures the pool surface profile, while an infrared sensor positioned at 30 degrees measures the thermal distribution. Both sensors are synchronized with the welding process through a common trigger signal.
Image processing: The visible light images are processed using edge detection algorithms to extract the pool width and surface contour. The infrared images are thresholded to identify the pool boundary based on temperature criteria (typically above 1500°C for steel welding).
Depth estimation: Pool depth is inferred from the relationship between measured surface width and thermal distribution, using empirical models calibrated from destructive cross-sectional measurements.
Real-time feedback: The processed parameters are fed back to the welding power source controller within a 50–100 ms control loop, enabling adjustment of current, voltage, or travel speed to maintain the desired pool geometry.

Key Findings and Performance Metrics

The study reports the following performance characteristics for the developed detection system:

Pool width measurement accuracy: ±0.3 mm for widths between 3–15 mm
Pool depth estimation accuracy: ±0.5 mm for depths between 1–8 mm
Measurement frequency: 100 measurements per second, providing 50–100 data points per typical weld bead
System response time: Less than 50 ms from measurement to control action
Signal-to-noise ratio: Greater than 20:1 under normal welding conditions
Welding parameter range: Current 50–300 A, travel speed 20–100 mm/min

The researchers demonstrated that the system can detect and respond to process disturbances such as arc wandering, filler wire misalignment, and base metal edge effects within a single control cycle, maintaining pool geometry within specified tolerances.

Engineering Applications and Quality Control Integration

The real-time molten pool monitoring system described in this study has direct applications in several areas of cladding and pressure vessel fabrication:

Overlay welding quality assurance: Monitoring pool geometry during cladding operations ensures consistent overlay thickness and dilution ratio, which are critical for the performance of corrosion-resistant and wear-resistant overlays.
Weld defect prevention: Real-time detection of pool anomalies such as excessive width (indicating porosity risk) or insufficient depth (indicating incomplete penetration) allows immediate corrective action before defects are locked into the solidified weld.
Process parameter optimization: The collected pool geometry data provides a quantitative basis for welding procedure qualification and optimization, reducing reliance on destructive testing for parameter validation.
Automated welding systems: Integration with robotic welding platforms enables adaptive control that maintains consistent weld quality across varying base material thicknesses, joint configurations, and production conditions.

Study Insights and Practical Considerations

This study addresses a practical gap in welding quality assurance: the transition from offline inspection to online monitoring. While the fundamental concept of optical pool monitoring has been researched since the 1980s, the integration of multiple sensing modalities with real-time processing and closed-loop control remains challenging in industrial environments. The collaboration between academic researchers and a manufacturing company is particularly noteworthy, as it ensures that the developed system addresses real production needs rather than purely academic questions.

For engineers implementing similar systems, several practical considerations should be noted:

Arc radiation interference: The intense visible and infrared radiation from the welding arc can saturate sensors and obscure pool features. Optical filters and sensor positioning must be carefully designed to balance signal capture with radiation protection.
Spatter and slag interference: Molten metal spatter and slag particles can temporarily obscure the pool surface, causing measurement errors. Robust signal processing algorithms are needed to distinguish true pool features from transient artifacts.
Calibration requirements: The system requires periodic calibration against known reference standards to maintain measurement accuracy over time, particularly as sensors age and optical components accumulate contamination.
Environmental factors: Ambient lighting, smoke, and fumes in the welding environment can affect sensor performance. Enclosed welding cells or local exhaust systems improve measurement reliability.

The study's contribution to the field is the demonstration that multi-sensor fusion can achieve the accuracy and speed required for real-time weld pool control in industrial TIG welding applications. This capability represents a significant step toward intelligent manufacturing in welding, where process parameters are continuously adjusted based on real-time feedback rather than fixed pre-programmed settings. Engineers in the cladding and pressure vessel industry should consider implementing such monitoring systems for critical applications where weld quality directly impacts safety and service life, recognizing that the initial investment in sensing technology is typically justified by the reduction in rework, scrap, and quality-related downtime.