Pattern Matching of Industrial Alarm Floods Using Word Embedding and Dynamic Time Warping

Dear Editor, This letter proposes a new pattern matching method based on word embedding and dynamic time warping (DTW) to identify groups of similar alarm floods. First, alarm messages are transformed into numeric values that represent alarms and also reflect the relationships between alarm occurrences. Then, similarities between numerically encoded alarm flood sequences are calculated by DTW and groups of similar floods are identified via clustering. The effectiveness of the proposed method is demonstrated by a case study with alarm & event data obtained from a public industrial simulation model.