Mixed-data multi-modelling for fault detection and isolation

0410790 - UTIA-B 20020004 RIV US eng J - Journal Article
Kárný, Miroslav - Nagy, Ivan - Novovičová, Jana
Mixed-data multi-modelling for fault detection and isolation.
International Journal of Adaptive Control and Signal Processing. Roč. 16, č. 1 (2002), s. 61-83. ISSN 0890-6327. E-ISSN 1099-1115
R&D Projects: GA ČR GA102/99/1564; GA MŠMT VS96063
Grant - others:EC(XE) IST9912058
Institutional research plan: CEZ:AV0Z1075907
Keywords : fault detection and isolation * mixed mixture models * quasi-Bayes adaptive estimation
Subject RIV: BC - Control Systems Theory
Impact factor: 0.540, year: 2002
http://library.utia.cas.cz/prace/20020004.ps

Early recognition/isolation of a faulty behaviour of a dynamic system is presented. Fault detection and isolation methods based on adaptive probabilistic models with multiple modes represent a theoretically well justified way of solution.The complexity problem is addressed here by employing an efficient quasi-Bayes estimation algorithm. It is directly applicable to the mixture of components created as products of factors belonging to the exponential family. The theory is illustrated by a simulation example.
Permanent Link: http://hdl.handle.net/11104/0130877