Title :
Using nonlinear black-box models in fault detection
Author_Institution :
Campus de Beaulieu, IRISA, Rennes, France
Abstract :
A method for fault detection is proposed using nonlinear black-box models. It is based on statistical tests derived from the local approach to change detection and the identification of black-box models. Partial physical knowledge, if available, can be combined with black-box models to handle the problem of over-parametrization
Keywords :
fault diagnosis; identification; monitoring; statistical analysis; fault detection; fault detection and isolation; nonlinear black-box models; over-parametrization; partial physical knowledge; statistical tests; Artificial neural networks; Condition monitoring; Equations; Fault detection; Mathematical model; Neural networks; Parametric statistics; Product safety; Production systems; Testing;
Conference_Titel :
Decision and Control, 1996., Proceedings of the 35th IEEE Conference on
Conference_Location :
Kobe
Print_ISBN :
0-7803-3590-2
DOI :
10.1109/CDC.1996.574396
