Research on the gearbox fault diagnosis based on SCS-BP neural network

Abstract

Neural network provides a new research method for equipment fault diagnosis with its inherent memory ability, self-learning ability and strong fault tolerance. Firstly, this paper studies the fault diagnosis and BP Neural Network and proposes an improved cuckoo search algorithm. Secondly, the research builds the model of fault diagnosis for equipment with CS-BP on the basis of equipment fault diagnosis characterized mathematical. Moreover, use a self-adaptive method to improve the cuckoo search algorithm. Finally, Case study is provided to illustrate the application of the proposed model. The model of improved CS-BP neural network has higher prediction accuracy and adaptability than the traditional BP neural network and has important application in the field of fault diagnosis.