Fault localization using neural networks and observers for autonomous elements
Por:
Benítez-Pérez H., Cárdenas-Flores F., Ortega-Arjona J.L., García-Nocetti F.
Publicada:
1 ene 2007
Resumen:
Fault detection and isolation (FDI) has become a useful strategy for determining fault appearance and on-line reconfiguration. However, unknown scenarios during on-line performance are still an open field for research. Different methods, such as knowledge-based techniques or analytical redundancy, have been followed. Nevertheless, both methods present inherent drawbacks for isolation. The present paper introduces a combined approach of model- and knowledge-based methods, using an autonomous element for isolation of unknown scenarios during on-line stage. The contribution is to integrate both methods to accomplish fault localization for unknown scenarios, based on previous information. Faults are constrained to certain bounded frequency response.
Filiaciones:
Benítez-Pérez H.:
Departamento de Ingeniería de Sistemas Computacionales y Automatización, IIMAS, UNAM, Apdo. Postal 20-726, Admón. No. 20, Del. A. Obregon, Mexico D. F., CP. 01000, Mexico
Cárdenas-Flores F.:
Departamento de Ingeniería de Sistemas Computacionales y Automatización, IIMAS, UNAM, Apdo. Postal 20-726, Admón. No. 20, Del. A. Obregon, Mexico D. F., CP. 01000, Mexico
Ortega-Arjona J.L.:
Departamento de Matemáticas, Facultad de Ciencias, Ciudad Universitaria, CP. 04510, México City, Mexico
García-Nocetti F.:
Departamento de Ingeniería de Sistemas Computacionales y Automatización, IIMAS, UNAM, Apdo. Postal 20-726, Admón. No. 20, Del. A. Obregon, Mexico D. F., CP. 01000, Mexico
|