Intelligent Alarm Flood Management for Power Systems

Alarm floods pose a challenge to the successful operation of transmission and distribution systems because the alarm rate is often higher than what operators can effectively manage. Such a high volume of alarms can distract operators from unrelated but relevant alarms that occurred during the flood. To mitigate this situation, we propose a method to group alarm floods into clusters and identify additional alarms that might be lost in the flood and ignored by the operator. To do so, three different clustering approaches are introduced and tested using historical alarm data from a transmission system operator (TSO). Importantly, the results are validated by the end user highlighting the applicability of the proposed alarm flood management tool in real world scenarios.