Automated Identification of Components of Feed Axes

Monitoring solutions like simulation models and digital twins enable methods like predictive maintenance, cutting down on idle time and improving the effectiveness and lifetime of machine tools. Creating these models for brown-field equipment is therefore sensible from both an ecological and economical view. Designing them, however, is a time-consuming process that requires a lot of expertise. Automating this has the potential to increase both the number and the quality of models. This paper presents an approach to automatically identify the parts of the feed axes of machine tools based on reference runs. For this, the control signals of exemplary machines are analyzed in order to develop a rule-based system to differentiate variations of parts. Furthermore, approximations of certain system parameters like gear ratios are determined. This constitutes a first step towards fully automated generation of functional models.