Prozessanalyse und -überwachung beim Metall-Schutzgasschweißen durch optische In-situ-Sensorsysteme

Purrio, Marion; Reisgen, Uwe (Thesis advisor); Schmitt, Robert Heinrich (Thesis advisor)

Aachen / Shaker Verlag (2016, 2017) [Book, Dissertation / PhD Thesis]

Page(s): 1 Online-Ressource (XV, 95, vii Seiten) : Illustrationen, Diagramme


Weld defects, such as spatters, incomplete fusion and weld anomalies in gas-metal arc welding result from a process which has not been optimally set or from variations of the process boundary conditions, for example, from the material composition, the weld preparation, the positioning and also from variations of the electrical parameters. The consequences are welding rejects or expensive weld finishing work. In order to counteract this, systems for process monitoring and process control are applied which provide machine-readable information about the actual state of the process via sensor systems. Previous approaches of process control were differentiated with regard to their positioning in the actual process into pre-, post- and in-situ methods. The aim of this work is to expand the hitherto existing possibilities of in-situ sensor systems and to utilise them for process control. In this work, methods are described which use image evaluation for information acquisition in gas-metal arc welding. To this end, the work is, besides introducing and concluding chapters, divided into two main parts: In the chapters 4 and 5, image evaluation as a means of inline sensor systems is used while events occurring during the process (offline) are detected and compared with recorded transient data. In doing so, the process is recorded with an image acquisition rate of up to 20000 Hertz and subsequently analysed. Here, it can be demonstrated that the droplet detachment in the pulsed process can be assigned to a clear characteristic of the voltage signal; but also the length of the wire stick-out can be detected in high-speed recordings and mapped onto the transient data. It is also possible to determine drop characteristics, such as the circumference or the trajectory and speed of the droplet, by means of these high-resolution recordings. In the next part of the work, image evaluation is treated directly as a component of inline sensor systems. Here, it is the aim to provide information which has been acquired from the molten pool directly for process monitoring and control. Chapters 6 and 7 give a detailed description of the methods which are used for detecting and evaluating the characteristics from images and also of the extent to which this is possible. The algorithms for the evaluation have, at that, been chosen in a way that a practical image acquisition and image evaluation rate can be obtained. The investigated characteristics are the wire electrode, the groove, the arc and the molten pool. From these objects, information about incorrect positioning of the torch in relation to the groove, the arc behaviour and the molten pool width can be obtained. A subsequent data and weld seam analysis also gives information about the effects of incorrect positioning.


  • REPORT NUMBER: RWTH-2017-08280