کنفرانس مهندسی برق ایران

صفحه اصلی / سی و چهارمین کنفرانس بین المللی مهندسی برق

Modeling Alarm Sequences Using Process Mining Techniques

نویسندگان :

Ali Norouzifar¹ Amir Neshastegaran² Iman Izadi³

1- Department of Electrical and Computer Engineering of Isfahan University of Technology 2- Department of Electrical and Computer Engineering of Isfahan University of Technology 3- Department of Electrical and Computer Engineering of Isfahan University of Technology

کلمات کلیدی :

Alarm Systems،Alarm Sequence Modeling،Process Mining،Process Discovery،Conformance Checking،Petri Nets

چکیده :

In industrial plants, faults often trigger characteristic sequences of alarms that reflect fault propagation through the system. Modeling such alarm sequences can therefore provide important insights into process behavior during abnormal situations. In this paper, a process mining–based approach is proposed to model alarm sequences associated with different fault scenarios. Process discovery methods are used to extract structured and interpretable models from historical alarm data. In particular, Inductive Mining and Evolutionary Tree Mining are employed to construct process tree and Petri net representations of alarm behavior. The discovered models are evaluated using conformance measures such as fitness, precision, and generalization. The approach is demonstrated using the Tennessee–Eastman benchmark process, and the resulting models can support downstream applications such as fault identification, root-cause analysis, and fault propagation analysis.