摘要：AbstractAn alarm flood is a commonly encountered problem in industrial alarm systems. It may distract operators from critical alarms and lead to plant failures. Thereby, addressing of alarm floods becomes an important task to achieve reliable alarm management. Some sequence alignment algorithms have been used to discover useful patterns of alarm floods from historical alarm data. In order to improve the computational efficiency of the existing algorithms, a fast sequence alignment approach is developed based on a basic local alignment search tool. It accelerates the computation using a seed-and-extend strategy. The priority information is incorporated so that the algorithm is more sensitive to alarms of higher priorities. Besides, the set based pre-matching is proposed to avoid unnecessary computation by excluding irrelevant alarm floods and alarm tags. Industrial case studies demonstrate that the proposed method significantly outperforms the existing algorithms.
