Fuzzy Diagnosis using Order of Magnitude Exaggeration

This paper proposes a fuzzy logic approach to allow qualitative symptoms that have been generated from order of magnitude qualitative simulation to be utilised as a fault detecting diagnostic system. The symptoms are interpreted as fuzzy rules and this allows numerical observations to be used as input to the symptoms and a ranking of faults as the output. The rule based approach results in a compact and fast (and fully testable) diagnostic system, by precompiling the results of off-line qualitative simulation and FMEA analysis. This paper outlines the technique which is being applied to automotive and aerospace electrical and fluid flow systems.