Seamless rings are used in several industrial sectors and are mainly produced by radial-axial ring rolling. To be able to detect form and process errors during the rolling process, several approaches with an extended sensor system (e.g. machine vision) have been made in recent years. It is a drawback of these approaches that the detection of an error requires it to be visible and to show a recognizable difference from a stable rolling process. To avoid a strong occurrence of errors during a rolling process, a different approach will be presented within this paper. To facilitate an early detection of an error, the process data of the rolling mill will be analysed synchronously to detect data constellations, which indicate the appearance of an error. With this information in an early stage of the error development, appropriate countermeasures can be initiated earlier compared to the manual reaction of an operator. Four rings have been rolled to show the potential of this approach exemplarily. The differences in the process data of all four rings and the relevant criteria to indicate the occurrence of an error will be presented within this paper.
展开▼