Application of Wavelet Neural Network with Particle Swarm Optimization Algorithm in Boiler Faults Diagnosis

摘要

In view of the main fault type in boiler steam water system, a variety of complex fault data are extracted. A wavelet neural network fault diagnosis based on particle swarm optimization algorithm is designed. The wavelet neural network constructed by three-layer wavelet neural network, is trained by particle swarm algorithm. By optimizing the weights factor, scale factor and shift factor wavelet neural network on particle swarm algorithm, the training speed of wavelet neural network is accelerated and the training accuracy is also improved. The simulation results show that the improved wavelet neural network algorithm is applied to the fault diagnosis of boiler, which can effectively eliminate the influence of redundant connection structure on network diagnostic ability and provide a new way for the boiler fault diagnosis.