Temporal analysis of event-discrete alarm data for improved manufacturing

Efficient response to machine failures is crucial for improving machine availability. In practice, the number of alarms occurring during the operation of manufacturing systems could overwhelm even experienced machine operators. Manufacturing execution systems (MES) therefore often allow for alarm data acquisition as well as its filtering and aggregation to reduce the occurring information overload. However, manual configuration of these systems may be cumbersome and time-consuming. Hence, we propose an event-discrete method for temporal alarm data analysis to support the configuration process.