Laser Vision Adaptive Fill Control for TIG Welding

Literature Overview

The paper by Fu Xibin and colleagues (2008), published in China Welding (Vol. 17, No. 4, pp. 1-6), presents an adaptive fill control system for TIG welding based on laser vision sensing. This work was conducted at the State Key Laboratory of Advanced Welding Production Technology at Harbin Institute of Technology. The system addresses a fundamental challenge in automated TIG welding: maintaining uniform weld fill when groove dimensions vary along the weld length due to machining tolerances, fit-up variations, or thermal distortion during welding.

Core Technical Points

The adaptive fill control system operates on a closed-loop feedback principle. A laser triangulation sensor measures the groove profile in real time as the welding process progresses. The measured groove geometry is compared with a pre-defined target profile, and the difference is used to adjust the welding parameters—primarily welding current and travel speed—to achieve the desired deposited metal volume and weld geometry.

The system hardware consists of four main components:

• A Modular Development Kit (MDK) serving as the real-time image capturing system, which processes the laser sensor data at high speed.
• A computer-based controller that runs the feature extraction algorithms and control logic.
• A D/A conversion card that interfaces the control signals to the welding power source.
• A DC TIG welding system as the controlled device, where the welding current is modulated in real time.

The feature extraction algorithm identifies the groove width, depth, and edge position from the laser sensor data. These features are then used in the control strategy to determine the required welding current adjustment. The control strategy employs a proportional-integral (PI) controller that adjusts the welding current based on the deviation between the actual and target groove dimensions.

System Architecture and Control Strategy

The following table outlines the system architecture and key control parameters:

The control strategy adjusts the welding current in proportion to the groove cross-sectional area. When the groove is wider than expected, the welding current is increased to deposit more metal. When the groove is narrower, the current is decreased to avoid excessive reinforcement. The travel speed is also modulated in coordination with the current to maintain a consistent heat input per unit length.

Experimental Results and Performance

The experimental results demonstrate that the adaptive fill control system achieves high stability, reliability, and accuracy in controlling the deposited metal volume. The weld formation quality meets relevant industry criteria, with consistent weld width, reinforcement height, and penetration across varying groove dimensions. The system effectively compensates for groove width variations of up to 0.5 mm and depth variations of up to 0.3 mm.

The feature extraction algorithm demonstrates good accuracy and robustness, with groove width measurement error within 0.15 mm and depth measurement error within 0.2 mm. The control system response time is less than 50 ms, which is sufficient to correct for groove variations at typical TIG welding travel speeds of 50-150 mm/min.

Engineering Practice Integration

For pipe welding applications, this adaptive fill control technology is particularly relevant to automated TIG welding of butt-weld joints in piping systems. The groove dimensions in pipe joints can vary due to bevel preparation tolerances, pipe ovality, and fit-up variations. The adaptive control system can compensate for these variations to produce consistent weld geometry, which is critical for meeting the stringent NDE acceptance criteria in pressure piping applications.

In the context of ASME B31.3 and API 1104 welding standards, consistent weld geometry is essential for achieving predictable NDE results. Variations in weld reinforcement and root geometry can lead to false indications in radiographic testing and ultrasonic testing. The adaptive fill control system helps ensure that the weld geometry remains within specified tolerances throughout the weld length.

Key Questions and Reflections

A practical consideration is the integration of the laser sensor with the welding torch assembly. The sensor must be positioned close enough to the weld zone to achieve accurate measurements, but far enough to avoid interference from the arc light and spatter. In practice, a fiber-optic laser sensor mounted on the welding torch with a standoff distance of 30-50 mm provides a good balance between measurement accuracy and operational practicality.

Another important aspect is the robustness of the control system against arc instabilities. The laser sensor measurements can be affected by arc light interference, and the control algorithm must include filtering and noise rejection capabilities to maintain stable operation. The use of a bandpass filter tuned to the groove feature frequency range, combined with a moving average filter, effectively removes high-frequency noise from the sensor data.

Study Insights and Implications

The laser vision adaptive fill control system represents a significant advancement in automated TIG welding technology. By enabling real-time adjustment of welding parameters based on actual groove conditions, the system achieves consistent weld quality across varying joint geometries. This technology is particularly valuable for automated pipe welding applications where joint geometry can vary significantly along the weld length. The combination of laser sensing, real-time feature extraction, and closed-loop control provides a robust solution for maintaining weld quality in production environments. The key to successful implementation lies in careful sensor positioning, robust algorithm design, and thorough validation under actual production conditions.