Visual Detection Method for No-Penetration Cladding Copper Molten Pool

Research Motivation and Technical Significance

No-penetration cladding (NPC) is a specialized weld overlay process designed to deposit a cladding layer onto a substrate without melting through to the opposite side, preserving the integrity of the base material. This process is particularly important for cladding copper or copper alloys onto steel substrates, where the large difference in thermal conductivity and melting point between the two metals makes traditional penetration welding impractical. The molten pool geometry during NPC cladding is complex and highly sensitive to process parameters, and visual monitoring of the molten pool is essential for real-time process control and defect prevention.

This literature presents a visual detection method for monitoring the copper molten pool during NPC cladding, using high-speed imaging and image processing techniques to extract key geometric features such as pool width, pool length, and pool surface temperature distribution. The ultimate goal is to develop a feedback control system that adjusts welding parameters in real time to maintain optimal molten pool geometry and prevent defects such as burn-through, insufficient fusion, and spatter.

Molten Pool Characteristics and Key Parameters

The copper molten pool during NPC cladding exhibits a distinctive elliptical shape with a leading edge that is hotter and more fluid than the trailing edge. The pool width typically ranges from 8–15 mm, and the pool length from 12–25 mm, depending on the welding current, travel speed, and wire feed rate. The surface temperature of the pool ranges from 1200°C (trailing edge) to 1450°C (leading edge), creating a significant thermal gradient that drives directional solidification.

The critical parameter for NPC cladding is the pool depth, which must remain below the substrate thickness to prevent penetration. For a typical 6 mm steel substrate, the maximum allowable pool depth is approximately 4–5 mm, leaving a 1–2 mm un-melted barrier that prevents burn-through. This depth control is inherently difficult to achieve without real-time monitoring, as the pool depth varies dynamically with process conditions.

Molten Pool Geometric Parameters

Visual Detection Methodology

The visual detection system employs a high-speed camera operating at 1000–2000 frames per second, positioned at a 30°–45° angle to the welding direction, with a focal length optimized for a field of view of 50–80 mm. The camera captures both visible light and near-infrared (NIR) images, which together provide complementary information about the molten pool geometry and temperature distribution.

Image processing algorithms extract the pool boundary by thresholding the intensity profile, followed by edge detection using the Canny algorithm. The pool width and length are measured as the maximum horizontal and longitudinal dimensions of the detected boundary, respectively. The surface temperature distribution is estimated from the NIR intensity using a calibrated emissivity model, with the copper emissivity assumed to be 0.65–0.75 in the 1000–1500°C range.

The detection accuracy is validated against thermocouple measurements and post-weld cross-sectional examination. The pool width measurement accuracy is within ±0.5 mm, the pool length accuracy within ±1.0 mm, and the temperature estimation accuracy within ±50°C. These accuracy levels are sufficient for real-time process control with a sampling rate of 10–20 Hz.

Defect Detection and Process Control Strategy

The visual detection system is designed to identify several key defect indicators in real time. An abnormally narrow pool width (below 6 mm) indicates insufficient heat input and risks incomplete fusion at the bond line. An excessively long pool (above 30 mm) indicates excessive heat input and risks burn-through. A sudden change in pool shape, such as a shift from elliptical to irregular, indicates process instability such as wire misalignment or shielding gas disruption.

A feedback control strategy is proposed where the detected pool parameters are compared against reference values, and the welding current, travel speed, and wire feed rate are adjusted proportionally to correct deviations. The control loop operates at a rate of 10–20 Hz, which is fast enough to respond to transient disturbances but slow enough to avoid oscillation. The control gains are tuned based on the dynamic response of the welding process, which was characterized through step-response testing.

Defect Indicators and Control Responses

Study Insights and Reflections

The literature demonstrates that visual monitoring of the molten pool is a practical and effective approach to improving the quality and consistency of NPC cladding processes. The key insight is that the molten pool geometry contains rich information about the welding process state, and that real-time extraction of this information enables proactive defect prevention rather than reactive inspection. The visual detection method bridges the gap between process monitoring and process control, providing the data foundation for closed-loop welding automation.

However, the literature also acknowledges the challenges of implementing this system in industrial settings. Copper is highly reflective, which creates specular highlights that can confuse image processing algorithms. The high temperature of the pool produces intense radiation that can saturate camera sensors. And the welding environment often includes spatter, smoke, and vibration that degrade image quality. Future work should focus on developing robust image processing algorithms that can handle these challenging conditions, integrating multi-spectral imaging for improved temperature measurement, and validating the system on a wider range of substrate thicknesses and cladding materials. The visual detection method represents a significant step toward intelligent process control in weld overlay applications, and its principles can be extended to other challenging cladding scenarios where real-time quality assurance is critical.