Application of Genetic Neural Network in Fault Diagnosis

Abstract

There are many kinds of fault diagnosis methods for analog circuit. Aiming at some shortcomings of BP neural network, optimizing BP neural network with genetic algorithm is studied in this paper. By optimizing the initial weights and thresholds of BP neural network, the network can obtain the best parameters in fast time, thus becoming the optimal network. The simulation results of an analog circuit show that the optimized neural network has a better global searching ability. It can quickly eliminate the solution with the much difference of the optimal solution and the convergence speed is improved. It overcomes the shortcomings of BP neural network to make inaccurate judgment of analog circuit and easy to fall into local minimum.