UNCERTAINTY AND APPROXIMATION IN MULTIMODEL DIAGNOSIS

Diagnosis of industrial plants is a complex task, usually performed by systems provided with the plant itself. However, these diagnostic sys tems are not reliable when the working conditions are different from t hose expected at design time. In this paper, we present OMISSYS (Oppor tunistic Model-based diagnosIS SYStem), a system that can diagnose fau lts in plants whose components are imperfectly described, and where da ta are affected by uncertainty and imprecision. OMISSYS exploits the a vailable knowledge and data to obtain a suboptimal diagnosis for the d etected fault. We represent imperfection using uncertain-fuzzy numbers , a formalism that can represent both uncertainty and imprecision. OMI SSYS was tested on realistic examples and it efficiently produced sati sfactory diagnoses. (C) EIsevier Science Inc. 1997.