FDI for a cooling system by using graphical approaches

In this paper, causal graphical models based approach for Fault Detection and Isolation (FDI) are adressed. Due to the structural and causal properties of a bond graph tool, a residual generation is based on analytical redundancy relations (ARRs) witch are obtained from the system behavior through different procedures in order to eliminate unknown variables (covering causal paths). Thus, the qualitative approach (temporal causal graph) is invoved to isolate faults by using the progressive monitoring for the refinement of the various fault hypotheses. The effectiveness of the proposed architecture is illustrated via simulation tests on a cooling system.