Pattern matching of alarm flood sequences by a modified Smith-Waterman algorithm

摘要

Alarm flooding is one of the main problems in alarm management. Alarm flood pattern analysis is helpful for root cause analysis of historical floods and for incoming flood prediction. This paper deals with a data driven method for alarm flood pattern matching. An alarm flood is represented by a time-stamped alarm sequence. A modified Smith-Waterman algorithm considering the time stamp information is proposed to calculate a similarity index of alarm floods. The effectiveness of the algorithm is validated by a case study on actual chemical process alarm data.