High Speed Algorithms for the Full Penetration Hole Detection in Laser Beam Welding Processes by Cellular Neural Networks
New visual algorithms for the control of keyhole welding processes are proposed in this paper. Keyhole welding allows obtaining highly focused laser beam, deep and slender weld seams and minimized heat affected zone at high feeding rates. These characteristics make the keyhole welding particularly suitable for several manufacturing processes, from automobile production to precision mechanics. Despite the improvement in welding technology, sophisticated methods of fault detection are not commonly used in commercially available equipments yet. Recent analyses of process images have revealed the possibility to adjust the laser power according to the detection of the so called full penetration hole. Due to the high welding dynamics, rapid physical movements of the full penetration hole can be observed. Therefore, robust closed loop control systems require fast real time image processing with frame rates in the multi kilo Hertz range. In the following, new Cellular Neural Network based strategies for the full penetration hole detection will be proposed. Such algorithms have been implemented in the Eye-RIS system v1.2 and frame rates up to 24 kHz have been reached. The best strategy in precision and time consumption has been already tested in real time applications and some experimental results will be also discussed.