Selection of the PWL excitation using SA algorithm with neural goal function

The paper presents a selection of the PWL test excitation for analog electronic circuit using the simulated annealing algorithm (SA) with the objective function based on self-organizing artificial neural network (SOM). The primary objective of the method is to detect any single parametric fault using signals measured at three outer accessible nodes of the system. Optimization is subjected to PWL input voltage which is fed to the input of the filter. The idea of a hybrid optimization method consists of SA algorithm (searches the space of solution) but the assessment is based on SOM analysis. It has been designed test stimuli for the filter Sellen-Key, and analyzed the effectiveness of this method for various parametric fault models. The simulation results indicate the high effectiveness of the proposed method.