Lehrstuhl für Fertigungstechnologie, Universität Erlangen-Nürnberg

Characterisation of CO2 Laser Beam Welding Process by Methods of Chaotic Dynamics and Neural Networks

Project Status: finished


In the project the characterization of a laser beam welding processes depending on the feed rate and harmonically modulated laser power is investigated. The final goal of the project is an in-process characterization of the penetration depth using plasma intensity signals. For this purpose, beside conventional methods, methods of nonlinear and chaotic time series analysis are used. With the aim to model the relations between characteristics of plasma intensity fluctuations and penetration depth a computer simulated neural networks will be applied. In the figure a development of process dynamics in dependence of feed rate is presented by means of spectrogram and coarse-grained entropy rate CER. Positions of CER?s local maxima indicate the possible transition regions between different dynamics of the process.

Letztes Update: 28.04.2010