Application of Genetic Algorithms to Analog Fault Diagnosis

This paper addresses itself to analog fault diagnosis by means of simulation-before -test approach, the so called dictionary approach. Original diagnostic system for single fault detection, location and identification is built, based on the Genetic Algorithm (GA) technique. The proposed system allows identification of the selected parametric faults at reasonable dictionary size and low computational effort. This feature is not present in classical dictionary. The system effectiveness has been verified by computational examples and the obtained results have confirmed usability of the GA technique to analog dictionary construction .