Hierarchical diagnosis strategy fora n auto motive electrical power generation system

This paper presents a hierarchical signal knowledge-based approach to fault detection and isolation (FDI) for an automotive electrical power generation system. The system we deal with is characterized by two main components: an alternator and a belt and it is subjected to constraints on the available measures (the alternator current is not available) and the computational power. The proposed approach is able to detect and isolate different faults combining, into a rule based logic, information carried by the measured signals. The hierarchical feature of the proposed diagnostic logic makes the implementation computationally efficient allowing the diagnosis algorithm to run in real time.