This paper introduces the GA（ genetic algorithm）-PSO （particle swarm optimization） algorithm, which can perform the GA operations of selection, crossover and mutation, and expounds the parameter optimization process applied to BP neural network. Then this GA-PSO-BP neural network is applied in fault diagnosis of generator units, and the simulation test has been conducted on rotor test-bed, which shows that GA-PSO-BP is superior to conventional BP and PSO-BP in training speed and diagnosis accuracy.
BP neural networkGA-PSO algorithmgenerator unitfault diagnosis
NETL
NSTL
维普
万方数据
