POSSIBILISTIC CAUSAL DIAGNOSIS: FROM SINGLE TO MULTIPLE FAULTS
S. Boverie** D. Dubois** O. de Mouzon* H. Prade**
* DGEI, INSA, 135 a. de Rangueil, 31077 Toulouse Cx 4, France
** IRIT, UPS, 118 r. de Narbonne, 31062 Toulouse Cx 4, France
*** Siemens VDO Automotive S.A.S., 1 avenue Paul Ourliac, BP 1149, 31036 Toulouse Cedex 1, France
A general approach to diagnosis based on fuzzy pattern matching is proposed, making use of consistency and inclusion-based indices in the setting of possibility theory. It first concentrates on single-fault detection, using crisp sets before fuzzy sets. Then the diagnosis is extended to multiple-fault detection, making use of information about normal values of attributes.
Keywords: Diagnosis, engine dynamometer, uncertainty, fuzzy sets, possibility theory
Session slot T-Tu-E10: Fault Diagnosis Design II/Area code 7e : Fault Detection, Supervision and Safety of Technical Processes

|