Self-optimizing focus technique (SOFT) – A new approach on the way to next generation cognitive laser cutting machines Available to Purchase

Within the framework of the Cluster of Excellence “Integrative Production Technology for High-Wage Countries” of RWTH Aachen University, Fraunhofer ILT is strongly engaged in the project “Integration of self-optimizing set-up, monitoring and control systems in production processes”. In this context, Fraunhofer ILT is focusing on next generation cognitive laser cutting machines. The work started with a detailed process analysis based on experimentation and theory, a definition of the most relevant and sensitive process parameters and the identification of a suitable concept for the process sensor system. A metamodel is the central, methodically integrative tool, in order to store, organize and make available theoretical and practical expert experience and knowledge for the cognitive decision routines to be developed. As an essential prerequisite, the correlation of applicable process parameters, detectable process image characteristics and achievable cut quality features has to be understood and represented within the metamodel. The most relevant process parameter, namely the focal position is selected as the optimizing setting parameter, owing to the difficulties in pre-setting its absolute value, owing to the time-dependent and parameter-dependent thermal behaviour of optical elements at multi-kW power levels, as well as owing to its serious influence on process efficiency, robustness and quality. A high-speed vision system - coaxially integrated in the beam path - delivers image data containing a measure of the optimization. Different levels of Self-Optimizing Focus Techniques (SOFT) will be accomplished. They are classified according to the knowledge-level and decision complexity to be managed by the “SOFT-Controller”. This varies from a machine assistant for focal position calibration in a set-up mode, up to a closed-loop in-process self-optimisation of cut quality based on cognitive decision routines. The developed functionalities are gradually integrated and qualified in a TRUMPF TruLaser 5030 laser cutting machine.