Visual Detection of Copper Molten Pool Images in Automatic TIG Hardfacing
Literature Overview and Sensing Approach
This 2004 paper by Wang Kehong and colleagues from Nanjing University of Science and Technology, published in the Journal of Welding, addresses a critical challenge in automatic TIG hardfacing of copper: how to obtain clear images of the molten pool for real-time process control. The authors proposed a near-infrared band imaging approach and designed two composite narrow-band filter systems based on different imaging mechanisms. Using a standard CCD sensor, they captured images of the copper molten pool during pulsed TIG hardfacing from the front side, and compared the results under four filter conditions: 1064 nm, 980 nm, 520 nm, and 405 nm. The key finding is that the near-infrared bands (1064 nm and 980 nm) provide clearer molten pool images than the visible bands, and that during the pulse base period, all four filter bands can produce clear images.
Filter System Design and Image Acquisition
The challenge of imaging a molten pool during TIG welding is the intense arc radiation, which overwhelms the weak radiation from the molten pool surface. The authors addressed this by using narrow-band filters to isolate specific wavelengths where the arc radiation is relatively weak compared to the molten pool radiation. In the near-infrared band, the arc plasma emits less radiation than in the visible band, while the molten copper pool emits significant radiation due to its high temperature. This contrast allows the CCD sensor to capture a clear image of the pool without saturation from the arc.
| Filter Wavelength | Band Type | Image Quality | Best Acquisition Period |
|---|---|---|---|
| 1064 nm | Near-infrared | Clear | Pulse base period |
| 980 nm | Near-infrared | Clear | Pulse base period |
| 520 nm | Visible | Moderate | Pulse base period |
| 405 nm | Visible | Moderate | Pulse base period |
The authors also noted that the pulse TIG arc base period is the best time to capture images. During the base period, the arc current is low, and the arc radiation is reduced, allowing the CCD sensor to capture a clear image of the molten pool without saturation. This is a critical insight for practical process control. In a pulsed TIG hardfacing process, the base period is the ideal window for image acquisition, and the control system can be synchronised with the pulse cycle to capture images at the optimal moment.
Molten Pool Characteristics and Depth Control
The authors analysed the characteristics of the copper molten pool and its relationship to penetration depth. In TIG hardfacing of copper, the high thermal conductivity of copper makes it difficult to achieve consistent penetration. The molten pool tends to be wide and shallow, and the penetration depth is sensitive to variations in heat input and arc length. The authors' visual detection system allows real-time monitoring of the molten pool width and shape, which can be used to control the penetration depth by adjusting the welding current or travel speed.
In my experience, controlling penetration depth in copper hardfacing is one of the most challenging aspects of the process. Copper's high thermal conductivity means that heat is rapidly conducted away from the weld pool, making it difficult to achieve a consistent fusion zone. The molten pool tends to spread out rather than penetrate deeply, and the penetration depth is highly sensitive to the arc length and the angle of the torch. A visual detection system that can monitor the molten pool in real time is invaluable for maintaining consistent quality, particularly in automated hardfacing processes where the operator cannot directly observe the pool.
Engineering Practice and Application Insights
This research has direct relevance to the hardfacing of copper components, such as continuous casting moulds, electrical contacts, and heat exchanger tubes. These components require a high-quality hardfacing layer with minimal dilution and consistent penetration. The visual detection system described in this paper can be integrated into an automated TIG hardfacing system to provide real-time feedback for process control. In my experience, the integration of visual sensing with automated hardfacing is a key step toward achieving consistent quality in production. The ability to monitor the molten pool in real time allows the control system to adjust process parameters dynamically, compensating for variations in heat input, arc length, and material properties.
The authors' finding that the near-infrared band provides clearer images than the visible band is particularly important for industrial applications. In a production environment, the welding arc is often accompanied by smoke, spatter, and other visual interference that can obscure the molten pool. Near-infrared imaging is less affected by these interferences than visible imaging, making it more robust for industrial use. The use of a standard CCD sensor, rather than a specialised infrared camera, also reduces the cost and complexity of the system, making it more practical for widespread adoption.
Key Questions and Reflections
One question this paper raises is the applicability of the visual detection system to other hardfacing processes, such as plasma arc hardfacing or laser hardfacing. The principles of near-infrared imaging and pulse-synchronised image acquisition are likely to be applicable to these processes as well, but the optimal filter wavelengths and acquisition timing may differ. In my experience, the molten pool characteristics vary significantly between processes, and the sensing system must be tailored to the specific process and material. Another question is the effect of the molten pool image on the control algorithm. The authors analysed the molten pool characteristics and their relationship to penetration depth, but they did not describe a complete closed-loop control system. The next step would be to develop a control algorithm that uses the molten pool image to adjust process parameters in real time.
Study Insights and Implications
This paper demonstrates the feasibility of using near-infrared imaging with a standard CCD sensor to obtain clear images of a copper molten pool during pulsed TIG hardfacing. The findings have direct implications for the automation of copper hardfacing processes, enabling real-time monitoring and control of the molten pool. For any engineer responsible for automating hardfacing processes, this paper is a valuable reference that highlights the importance of proper filter selection and pulse synchronisation for image acquisition. The use of near-infrared imaging is particularly effective for high-temperature materials like copper, where the arc radiation is intense and the molten pool radiation is relatively weak. This research represents a significant step toward fully automated hardfacing processes with consistent quality and minimal operator intervention.
Zhuojin Pipe Fitting Co., Ltd