Signal interpretation for monitoring and diagnosis, a cooling system testbed

This paper discusses a method for fault detection and isolation (FDI) in co ntinuous dynamic systems. A key aspect of this approach is the coupling of a qualitative diagnosis engine and a monitoring system that computes symbol ic feature values through a signal-to-symbol transformation on the continuo usly sampled measurement data, Signal analysis techniques with a sound stat istical basis are employed to generate reliable symbolic data. The methodol ogy is evaluated on the diagnosis of engineered faults in the cooling syste m of an automobile engine that has been instrumented with temperature and p ressure sensors. Results show the interdependency between modeling for diag nosis and the feature extraction system.