Visual Detection of Copper Weld Pool in No-Penetration Overlay Welding

Literature Overview

The paper by Xu Yuelan and colleagues from Nanjing University of Science and Technology, published in the Journal of Nanjing University of Science and Technology in 2005, investigates the visual detection of copper weld pools during tungsten inert gas (TIG) welding without penetration. The study analyzes the imaging principles for weld pool detection, proposes near-infrared band imaging for copper weld pools, and designs composite narrow-band filter systems to achieve clear weld pool images. The research reveals that copper weld pools are more difficult to image clearly than steel weld pools in the near-infrared band, a finding with significant implications for process monitoring and control.

Fundamentals of Weld Pool Visual Detection

Weld pool monitoring through visual imaging is a critical technology for welding process control and quality assurance. The weld pool geometry, temperature distribution, and flow patterns directly influence weld bead shape, penetration depth, and defect formation. Visual detection systems capture images of the weld pool in real-time, enabling closed-loop control of welding parameters such as travel speed, wire feed rate, and arc length.

For no-penetration overlay welding, the weld pool exists entirely on the surface of the base material without penetrating through it. This condition is essential for overlay applications where the base material must remain intact, such as repair welding of thin-walled components or surface hardening of existing parts. The weld pool in this scenario is characterized by a shallow, wide geometry with limited depth, which presents unique challenges for visual detection due to the reduced thermal contrast and smaller imaging area.

Imaging Parameters and Filter Design

The selection of the near-infrared band for imaging is based on the spectral emission characteristics of the TIG arc and weld pool. The TIG arc emits strong radiation across the visible and near-infrared spectrum, with peak emission intensity in the near-infrared region for copper materials. The near-infrared band offers a balance between sufficient signal intensity and reduced interference from arc radiation, which is more intense in the visible band.

The composite narrow-band filter system is designed to selectively transmit the wavelength range where the weld pool emits the strongest radiation while blocking the surrounding arc radiation. The filter design incorporates multiple narrow-band elements to create a composite passband that optimizes the signal-to-noise ratio. The specific filter configuration is tailored to the material being welded, with copper requiring different filter parameters than steel due to differences in emissivity and spectral emission characteristics.

Spectral Analysis and Copper Weld Pool Characteristics

The spectral analysis of the TIG arc reveals distinct emission peaks associated with the arc plasma and the molten pool. For copper materials, the arc spectrum shows strong emission lines from copper atoms and ions in the visible and near-infrared regions. The weld pool of copper, however, exhibits a broader and less intense emission spectrum compared to steel, which is attributed to the lower emissivity of molten copper and the different temperature distribution within the pool.

The emissivity of molten copper is approximately 0.6 to 0.7 in the near-infrared band, compared to 0.8 to 0.9 for molten steel. This lower emissivity results in weaker thermal radiation from the copper weld pool, making it more difficult to distinguish from the surrounding base metal and arc radiation. The temperature of the copper weld pool is approximately 1300 to 1500 degrees Celsius, which is lower than the steel weld pool temperature of 1500 to 1800 degrees Celsius, further reducing the radiation intensity according to the Stefan-Boltzmann law.

Comparative Imaging Results

The imaging results demonstrate that achieving a clear image of the copper weld pool requires more aggressive filtering and longer exposure times compared to steel. The pool boundary in copper welds is less distinct due to the gradual temperature gradient at the pool edge and the lower contrast between the molten pool and the solid base metal. This makes automated pool tracking and geometric measurement more challenging for copper overlay welding applications.

The study also identifies that the observation window, defined as the angle between the camera axis and the weld pool surface normal, significantly affects image quality. For copper welds, the optimal observation angle is closer to the normal direction (0 to 15 degrees), whereas for steel welds, a wider range of angles (0 to 45 degrees) can produce acceptable images. This difference is attributed to the specular reflection characteristics of molten copper, which is more reflective than molten steel in the near-infrared band.

Process Control Implications

The difficulty of imaging copper weld pools has direct implications for the development of automated welding control systems. Pool tracking algorithms that rely on image processing to determine pool geometry and position are less reliable for copper materials, requiring more sophisticated image processing techniques and potentially additional sensing modalities.

For no-penetration overlay welding of copper components, the process control strategy must account for the limited pool visibility. Alternative monitoring approaches such as infrared thermography, acoustic emission, or current-voltage waveform analysis may need to be combined with visual detection to achieve adequate process control. The multi-sensor fusion approach provides redundant information that compensates for the limitations of any single sensing modality.

The research findings contribute to the broader understanding of material-specific challenges in welding process monitoring. The spectral and emissivity differences between materials must be considered when designing welding monitoring systems, and the assumption that a monitoring system developed for one material will perform equally well for another material is not valid. Material-specific calibration and optimization are essential for reliable process control.

Engineering Application and Future Directions

The practical application of this research lies in the development of reliable visual monitoring systems for copper overlay welding applications. Copper overlay welding is used in various industrial applications including electrical contact repair, heat exchanger tube repair, and anti-galling surface coatings. The ability to monitor the weld pool in real-time is essential for ensuring consistent weld quality and preventing defects such as undercut, excess penetration, or incomplete fusion.

The research also highlights the need for material-specific filter design in welding monitoring systems. Commercial welding monitoring systems typically use broadband filters optimized for steel welding, which may not provide adequate performance for copper or other non-ferrous materials. Custom filter design based on the spectral emission characteristics of the specific material is a critical requirement for reliable monitoring.

The findings of this study underscore the importance of fundamental research in welding process monitoring. The spectral and emissivity properties of different materials create distinct challenges that require targeted solutions rather than one-size-fits-all approaches. For welding engineers developing automated welding systems, understanding these material-specific factors is essential for achieving reliable and robust process control.