DocumentCode :
2766229
Title :
Diagnosing hybrid dynamical systems: fault graphs, statistical residuals and Viterbi algorithms
Author :
Basseville, Michéle ; Benveniste, Albert ; Tromp, Laurent
Author_Institution :
IRISA, Rennes, France
Volume :
4
fYear :
1998
fDate :
16-18 Dec 1998
Firstpage :
3757
Abstract :
We address the problem of model-based fault diagnosis for hybrid numeric/symbolic dynamical systems. The objective is the systematic design of fault diagnosis algorithms, based on fault graphs provided by reliability analysis and on-board sensor data processing. We propose an approach, based on HMM ideas and a Viterbi algorithm. Experimental results of the proposed approach for a simulated mechanical engineering example are reported. Extensions for taking into account a discrete controller and fault concurrence are discussed
Keywords :
Petri nets; fault diagnosis; hidden Markov models; reliability theory; statistical analysis; Viterbi algorithms; discrete controller; fault concurrence; fault graphs; hybrid dynamical systems; model-based fault diagnosis; reliability analysis; simulated mechanical engineering example; statistical residuals; Costs; Data processing; Failure analysis; Fault detection; Fault diagnosis; Hidden Markov models; Mechanical sensors; Monitoring; Sensor systems; Viterbi algorithm;
fLanguage :
English
Publisher :
ieee
Conference_Titel :
Decision and Control, 1998. Proceedings of the 37th IEEE Conference on
Conference_Location :
Tampa, FL
ISSN :
0191-2216
Print_ISBN :
0-7803-4394-8
Type :
conf
DOI :
10.1109/CDC.1998.761804
Filename :
761804
Link To Document :
بازگشت