Ultrasonic-driven adaptive control of robotic plasma arc cutting for bevel applications

In heavy industries like oil and gas, and shipbuilding, maintaining process quality is challenging. These sectors face inconsistent manual procedures and a shortage of skilled operators regarding thermal cutting and bevelling for welding preparation tasks. Manual fitting and repetitive quality control modifications, especially during thermal cutting, significantly increase time consumption and hinder productivity. Traditional thermal cutting methods are prone to human error, resulting in inconsistent cut quality, and demand high expertise leading to variability in cut precision, increased rework, and material wastage. The objective of this work is to address these challenges by introducing real-time ultrasonic sensing into a robotic plasma cutting control system to automate the steel plate bevelling process. The ultrasonic sensor enables the system to dynamically adapt to variations in steel plate thickness before cutting, ensuring precise and consistent results. The solution begins by presenting an automated method for measuring thickness and computing bevel distance per sample. Secondly, it proposes adaptive adjustments to cutting parameters per sample, leveraging the ultrasonic sensor data to enhance accuracy and reduce the need for manual intervention. Finally, the approach introduces adaptive robotic path generation for cutting and utilizing real-time ultrasonic sensor data to optimize cutting paths. The outcome of this study is the successful development and validation of an adaptive robotic plasma cutting system for steel plate bevel applications, which leverages real-time ultrasonic sensor data to automate the parameter input process and robotic motion planning, demonstrating improved accuracy and efficiency compared to traditional approaches. The results demonstrate that ultrasonic-driven robotic cutting significantly reduces the average error cut percentage to 4.47% with deviations ranging from 0.13 to 0.23° for the bevel angle and 14.27% with deviations between 0.02 and 0.05 mm for root face deviation, compared to the standard cutting approach which has an average error of 18% with deviations ranging from 0.10 to 0.38 mm and 77.1% with deviation between 0.48 to 0.90°, respectively. This paper highlights the benefits of using advanced sensing technology, particularly ultrasonic sensors, to automate plasma bevel cutting for metal plates in the steel fabrication and welding sectors.