Title :
A novel fault prediction technique using model degradation analysis
Author :
Lennox, B. ; Rutherford, P. ; Montague, G.A. ; Haughin, C.
Author_Institution :
Newcastle upon Tyne Univ., UK
Abstract :
This paper presents two practical applications where artificial neural networks have been used to solve difficult process engineering problems. Firstly, the ability of artificial neural networks to provide an accurate process model of a vitrification process is demonstrated on real process data. Vitrification is the process which encapsulates highly active liquid waste in glass to provide a safe and convenient method of storage. The second application again employs artificial neural networks, this time they are applied in a novel way in which they are used to capture non-linear system characteristics and then recalled to provide a means of detecting imminent failure of a vessel used in the vitrification process
Keywords :
fault diagnosis; identification; neural nets; process control; artificial neural networks; fault prediction technique; highly active liquid waste; imminent failure detection; model degradation analysis; process engineering problems; vitrification process; Artificial neural networks; Chemical processes; Degradation; Fault detection; Glass; Power system modeling; Predictive models; System identification; Systems engineering and theory; Vitrification;
Conference_Titel :
American Control Conference, Proceedings of the 1995
Conference_Location :
Seattle, WA
Print_ISBN :
0-7803-2445-5
DOI :
10.1109/ACC.1995.532208
