An alarm flood pattern matching algorithm based on modified BLAST with leveshtein distance

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 introduces a data driven method for alarm flood pattern matching. A modified BLAST algorithm with Levenshtein distance is proposed to discover similar alarm floods. A quadruple tank process is simulated to test the effectiveness of the proposed algorithm. The results show that the algorithm is quick and accurate.