Camera based in-process control for laser welding enables flexible image processing which allows the adaption of the system to different processes and quality features. A closed loop control system with a Cellular Neural Network Camera was implemented into a laser welding machine. The system is surveying the contour of the full penetration hole with a frame rate of over 10 kHz for both, acquisition and evaluation of area images. As a result the system reaches and holds the full penetration state automatically. This paper shows the latest experimental results including the extension to direction independent weldings.

Today, image processing using coaxial camera setups is used to monitor the quality of laser material processes such as laser welding, cutting or ablation. This article shows the potentials of a sensing system for the next step: A closed loop control of a full penetration keyhole welding process.With “Cellular Neural Networks” (CNN) it is possible to integrate processor elements in the electronic circuitry of CMOS cameras resulting in a Single-Instruction-Multiple-Data (SIMD)-architecture on the camera chip itself. Such pixel parallel systems provide extremely fast real-time image processing. This allows it to employ algorithms for image processing which are widely independent of the exact process parameters, reducing the adaption effort for different applications, compared to photo diode systems.A closed loop control system was implemented into a commercially available laser welding machine. This system uses a CNN based camera surveying the contour of the full penetration hole with a control frequency of up to 14 kHz for linear weldings. As a result the system reaches and holds the full penetration state automatically. A more detailed description of the control system and the used algorithms is given in [1].This paper presents achievable welding results in scanner-based remote welding processes, with a new direction independent algorithm. To demonstrate the capabilities of the closed loop control, welding experiments with changing process conditions were carried out.

The latest generation of high-brightness disk and fiber lasers allows deep penetration welding with a weld seam width of a few 10µm and an aspect ratio of up to 15. Welding speeds exceeding 60m/min are achieved. These properties are very promising for high-productivity joining of small parts like thin sheet and fine tubes with a thickness of up to 1 mm as frequently used in the electronic, automotive and medical industries.For a reliable production not only the machinery and tools like axes and clamping devices have to be adapted to the high welding speed and to the small weld dimensions, but also the process monitoring system sampling rate must be increased in order to get enough information per weld seam length.This paper presents a multi sensor system which is capable of monitoring the achieved weld quality of fine tubes and of detecting typical welding errors such as humping and the formation of blowholes.

Today, image processing using coaxial camera setups is used to monitor the quality of laser material processes such as laser welding, cutting or ablation. This article proposes a sensing system for the next step: Using image based quality features to form an instant feedback signal in order to maintain the process in the desired state.The key component of the system is a camera based on Cellular Neural Networks (CNN). This technology enables real time image processing which is necessary for a robust feature analysis in highly dynamic laser processes. It is used to control the laser power of a welding system by the contour of the full penetration hole. Compared to conventional systems, the frame rate for both, acquisition and evaluation rises from about 1 kHz to 14 kHz.This paper presents the measurement technology including algorithms and some welding results which demonstrate that the full penetration state is maintained even under rapidly changing process conditions such as steps in material thickness or variation of speed. An in-depth discussion on welding results is found in [1].

The growing number of laser welding applications from automobile production to micro mechanics require fast and reliable process control systems. The high process dynamics in time, space and intensity, especially in scanner based remote welding or high speed micro welding, demand extremely fast and spatially resolved in-process control systems to create closed loop control for error prevention and correction.Today’s conventional micro processor based image processing architectures (as used for example in [1]) are not able to provide the high frame rates needed for the real-time closed loop control of high speed laser welding.With “Cellular Neural Networks” (CNN) it is possible to implement Single-Instruction-Multiple-Data (SIMD)-architectures in the electronic circuitry of each pixel of the camera chip itself in order to produce a so called Focal Plane Processor (FPP). Such pixel parallel systems provide extremely fast real-time image processing. With these new CNN-cameras it is now possible to implement a camera based high speed in-process control system for laser welding that enables closed loop control of various quality features.With a multi modal diagnostic approach we were able to identify direct and explicit image attributes for a variety of quality features as a base for the process control. It could be shown that closed loop control of the “full-penetration” quality feature is possible with frame rates of 10 kHz and beyond with a CNN-camera system.