Beschreibung
The production of hybrid components involves a long process chain, which leads to high investment costs even before machining. To increase process safety and process quality during finishing, it is necessary to provide information about the semi-finished parts geometry for the machining process and to identify defect components at an early stage. This paper presents an investigation to predict variations in dimension and cavities inside the material during cross-wedge rolling of shafts based on measured tool pressure. First, the process is investigated with respect to the variation in diameter for three roll gaps and two materials. Subsequently, features are generated from the hydraulic pressures of the tools and multi-linear regression models are developed in order to determine the resulting diameters of the shaft shoulder. These models show better prediction accuracy than models based on meta-data about set roll gap and formed material. The features are additionally used to successfully monitor the process with regard to the Mannesmann effect. Finally, a sensor concept for a new crosswedge rolling machine to improve the prediction of the workpiece geometry and a new approach for monitoring machining processes of workpieces with dimensional variations are presented for upcoming studies.
