A fuzzy logic system for analog fault diagnosis

In this paper, a fuzzy logic system for analog fault diagnosis is introduced. The ability of fuzzy logic to encode structured knowledge in a numerical framework is exploited. Circuit measurements are properly fuzzified. Fuzzy rule bases are constructed such that the behaviour of the circuit under test in both fault and fault-free situations is characterized. The outputs of the fuzzy rule bases are then defuzzified to recover crisp values for circuit elements. A new fault diagnosis procedure that utilizes the proposed fuzzy logic system is proposed. An illustrative example to demonstrate the potential of this procedure in fault isolation is presented.