Fuzzy based time domain analysis approach for fault diagnosis of analog electronic circuits

Abstract

A fuzzy logic system for the diagnosis of single and multiple faults in analog circuits is proposed in this paper. The simulation-before-test (SBT) approach has been employed. Given the description of the circuit-under-test (CUT), the step response is simulated and consistent measurements of the response parameters are made following which a comprehensive fault dictionary is compiled. The capability of the fuzzy logic paradigm to reorganize structured data into a numerical framework has been capitalized upon. Fuzzy rule bases are built to typify the behavior of the circuit under fault free and several faulty conditions. The adoption of the time response parameters results in a fault dictionary that is free from masking and hence augments accuracy in fault isolation. An illustrative example is presented to demonstrate the applicability of this method.