Model-free fault detection: application to Polymer Electrolyte Fuel Cell system
Résumé
A novel fault detection method is presented in this paper. The proposed
method can be considered as an extension of the model-free control which
is based on an ultra-local model. The model-free controller has
demonstrated a great ability to perform the control tasks despite the
presence of a fault. This undeniable advantage can mask the effect of a
fault resulting in an abnormal degradation impacting one of the
components ensuring the control loop or the system itself. The proposed
method allows coping with this type of event rapid coping. The basic
idea of the diagnostic method is to reconstruct the measured output of
the system from the ultra-local model used for system control. The
estimated output is compared with the measured one to generate a
residual which is employed as the fault indicator. An experimental
validation of the polymer electrolyte membrane fuel cell (PEMFC) system
is carried out in real time to detect a a system fault that results in
membrane flooding. The main advantage of the method is to be able to
simultaneously control the system and detect faults that may affect it
without needing an accurate knowledge of the mathematical/physical
system model. The results obtained are very promising for real-time
diagnosis of the polymer electrolyte membrane fuel cell systems.
Origine | Fichiers produits par l'(les) auteur(s) |
---|