Fault diagnosis of chemical processes utilizing signed directed graphs-improvement by using temporal information

Author

Shiozaki, J. ; Shibata, B. ; Matsuyama, H. ; O´shima, E.

Author_Institution

Yamatake-Honeywell Co. Ltd., Tokyo, Japan

fYear

1988

fDate

25-27 May 1988

Firstpage

461

Lastpage

466

Abstract

A fault-diagnosis algorithm using a signed directed graph as a model of the system is presented for the real-time diagnosis of failures. The accuracy of the algorithm has been improved to select candidates which are the most likely to be the real origin of failure, utilizing the time when the measured variable begins to show abnormality as the representation of the dynamic characteristic of the measured variable. The accuracy and speed of the improved algorithm are examined by its application to data obtained in experiments of fault diagnosis in tank-pipeline systems.<>

Keywords

chemical engineering computing; directed graphs; failure analysis; process computer control; chemical processes; dynamic characteristic; fault-diagnosis algorithm; measured variable; real-time diagnosis; signed directed graph; tank-pipeline systems; temporal information; Chemical processes; Data structures; Diagnostic expert systems; Fault diagnosis; Fault trees; Production; Real time systems; Sensor phenomena and characterization; Sensor systems; Tree data structures;

fLanguage

English

Publisher

ieee

Conference_Titel

Artificial Intelligence for Industrial Applications, 1988. IEEE AI '88., Proceedings of the International Workshop on

Conference_Location

Hitachi City, Japan

Type

conf

DOI

10.1109/AIIA.1988.13332

Filename

13332