Premium
A stylized trend analysis approach for process monitoring and fault diagnosis
Author(s) -
Fang Yan,
Rasel M.A.K.,
Richmond Peyton C.
Publication year - 2018
Publication title -
process safety progress
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.378
H-Index - 40
eISSN - 1547-5913
pISSN - 1066-8527
DOI - 10.1002/prs.11951
Subject(s) - stylized fact , process (computing) , fault (geology) , reliability engineering , engineering , fault tree analysis , fault detection and isolation , real time computing , computer science , actuator , electrical engineering , seismology , economics , macroeconomics , geology , operating system
A stylized trend analysis was developed to identify recurring or standard operating procedures (SOP) such as equipment switching or maintenance from plant process history data to detect faults and near misses due to operator errors or equipment failures. Trend based fault detection and diagnosis systems have not been implemented widely in the chemical process industries because known fault scenarios are typically required. Known fault scenarios were not required for the stylized trend analysis because faults and near misses were identified by comparing trends with normal SOP transition responses. Large volumes of historical data were processed automatically allowing corrective action to be taken prior to an incident. The stylized trend analysis was demonstrated to detect unsteady‐state transition faults on historical distributed control system data simulated for a continuous process ethylene plant dryer switching operation. © 2017 American Institute of Chemical Engineers Process Process Saf Prog 37:411–418, 2018