Research on annealing genetic hybrid optimization of test excitation in nonlinear analog circuit

The intelligent fault diagnosis method based on neural network is one of the most practical diagnosis methods in nonlinear analog circuit, the selection of appropriate test excitation signal can improve the diagnosis accuracy. The paper studies fault diagnosis excitation optimization method based on simulated annealing and genetic hybrid algorithm. In nonlinear analog circuit fault diagnosis takes Volterra kernel of low order as feature, the parameter selection of the multi-frequency sinusoidal excitation signal is taken as optimal problem, in case of same excitation, the lumped Euclidean distance of various fault states feature vector is taken as fitness function, the search method of optimum excitation signal parameter combining simulated annealing algorithm and genetic algorithm is studied. The optimization scheme of annealing genetic hybrid algorithm is presented and the practical examples are also given to verify the method in this paper.