Visual Detection of Molten Copper Pool in Automatic TIG Surfacing

Literature Overview

This paper by Wang Kehong and colleagues from Nanjing University of Science and Technology, published in Welding Journal (2004, Vol. 25, No. 4, pp. 27-30), presents a visual detection system for monitoring the molten copper pool during automatic TIG surfacing of copper deposits. Funded by the Jiangsu Provincial High-Tech Project (Grant No. BG2002021), this work addresses the critical challenge of optical monitoring in copper welding applications, where the high reflectivity and low emissivity of copper pose significant difficulties for conventional imaging systems.

Technical Challenge of Copper Pool Imaging

Copper and copper alloys present unique challenges for optical monitoring during welding. The high electrical and thermal conductivity of copper results in low emissivity in the visible and near-infrared spectrum, making it difficult to obtain clear images of the molten pool. Additionally, the intense arc radiation during TIG welding creates a strong background signal that can overwhelm the relatively weak signal from the copper pool. The researchers addressed these challenges by developing a compound narrow-band filter system that selectively transmits specific wavelengths while rejecting arc radiation.

Filter System Design and Comparison

The study designed two compound narrow-band filter systems based on different imaging mechanisms and tested four filter wavelength conditions: 1064 nm, 980 nm, 520 nm, and 405 nm. The imaging was performed using a standard CCD sensor positioned at the front of the welding zone during pulsed TIG surfacing of copper.

The 980 nm filter condition produced the clearest copper pool images, providing the best signal-to-noise ratio for pool monitoring. During the base current period of pulsed TIG welding, all four filter wavelengths were capable of producing clear pool images, which is an important finding for practical implementation. The peak current period, however, required the near-infrared filters to overcome the intense arc radiation.

Pulsed TIG Surfacing Process Characteristics

The pulsed TIG surfacing process was employed to control the molten pool geometry and penetration depth. The pulse parameters included base current, peak current, pulse frequency, and pulse width. The base current period maintains the pool in a molten state while the peak current period provides additional energy for penetration and alloy mixing. The visual detection system was designed to monitor the pool during the base current period, when the arc radiation is lower and pool imaging is more feasible.

The relationship between pool characteristics and penetration depth was analyzed through both imaging and macrograph examination. The pool width, length, and shape obtained from the visual images correlated with the actual penetration depth measured after welding. This correlation is essential for developing closed-loop control systems that adjust process parameters in real time based on pool monitoring feedback.

Engineering Significance for Surfacing Applications

The visual detection system described in this paper has direct application to automated surfacing processes where consistent pool geometry and penetration depth are critical. In the surfacing of copper deposits on steel substrates, such as for electrical contact applications or corrosion-resistant linings, the penetration depth must be carefully controlled to ensure adequate metallurgical bonding without excessive dilution of the copper deposit. The visual monitoring system provides a non-contact, real-time method for assessing pool conditions and adjusting process parameters accordingly.

The compound narrow-band filter approach is particularly valuable because it uses a standard CCD sensor rather than requiring specialized infrared cameras. This reduces system cost and complexity while maintaining adequate image quality for process monitoring. The 980 nm filter window is recommended as the primary imaging wavelength for copper pool monitoring, with the 1064 nm filter as a backup option.

Key Reflections and Practical Considerations

The study demonstrates that near-infrared imaging is the most effective approach for copper pool monitoring, which is consistent with the fundamental optical properties of copper. Copper has a relatively high emissivity in the near-infrared region compared to the visible region, making it more suitable for passive imaging. The pulsed TIG process was selected because the periodic variation in arc power creates windows of opportunity for imaging during the base current period.

For practical implementation, several factors must be considered. The imaging system must be shielded from ambient light and arc spatter, and the optical path must be maintained free of contamination. The filter system requires periodic cleaning and alignment to maintain optimal performance. The CCD sensor must be positioned at an appropriate angle and distance to capture the pool geometry without being damaged by arc radiation or spatter.

The correlation between pool image characteristics and penetration depth provides the foundation for developing automated process control algorithms. By tracking pool width, length, and shape in real time, the system can detect deviations from the target pool geometry and adjust current, voltage, or travel speed to maintain consistent weld quality. This approach is particularly valuable for long automated surfacing runs where manual monitoring is impractical.

Study Insights and Reference Value

This paper makes a significant contribution to the field of welding process monitoring by demonstrating that standard CCD sensors combined with compound narrow-band filters can achieve adequate copper pool imaging during TIG surfacing. The identification of 980 nm as the optimal imaging wavelength provides a clear design guideline for future monitoring system development. The correlation between pool geometry and penetration depth establishes the basis for closed-loop process control. For engineers working on automated surfacing of copper and copper alloys, this study provides a practical and cost-effective approach to visual process monitoring that can improve deposit quality consistency and reduce reliance on destructive testing for quality verification.