<scp>Hidden Markov</scp> model based approach for diagnosing cause of alarm signals

When a fault occurs in process, it slowly propagates within the system and affects measurements triggering sequence of alarms control room. The operators are required to diagnose cause take necessary corrective measures. idea representing alarm as propagation path using is explored. A diagnoser based on hidden Markov model built identify signals. proposed approach applied an industrial case study: Tennessee Eastman process. results show that successful determining probable generated with high accuracy. was able accurately, even when tested short sub-sequences. This allows for early identification faults, providing more time operator restore normal operation.