PROCESS MINING IN PRODUCTION SYSTEMS

Due to the increasing automation and integrity of today’s productions systems, thousands of alarms are generated every day in the more and more complex process control units. The core concept of our work is the investigation of the application possibilities of the Process Mining Framework (PROM) for the mining of industrial alarm management databases. The taskoriented formalisation of input datasets in the form of XES files is presented. The applicability and effectiveness of the different plug-ins are shown in terms of the analysis of the alarm and event-log database of an industrial delayed-coker plant. The results illustrate that the presented tools are suitable for the tracking of activities at different levels of the hierarchy, for the detection of the spillover effect of different malfunctions and the detection of the changes in the operating conditions. Index Terms—Alarm management, Process mining, Fault detection and diagnosis, Operator effectiveness

János Abonyi, Gyula Dörgő, IEEE, 2019