Processing Signal Parameters Based Fuzzy Inference System Classifier for Analog Circuit Single Parametric Faults

摘要

This paper presents a diagnostic method for singular parametric faults in analog circuits based on the Fourier coefficients and Distortion rate. The Simulation Before Test (SBT) approach has been used. In the proposed method, the DC component, the distortion rate and the fundamental amplitude extracted from Fourier analysis of the circuit output voltage signal have been selected to determine the circuit state conditions. The selection of these parameters provide a better characterizing mean of the circuit state conditions and contribute in the faults dictionary construction under faulty and fault free circuit conditions. The faults in concern are of singular parametric fault type that corresponds to a deviation from ±10 to ±50% with a 10% step variation of the circuit component nominal value. The faults classification is carried out using the Fuzzy Inference System (FIS) where the input data are the aforementioned Fourier parameters. The proposed method is illustrated through the Sallen-Key band pass filter test circuit. The obtained results show that the faults ambiguity problem being totally solved which confirm the efficiently of the proposed method in the diagnosis of the singular parametric faults in analog circuits.