An Intelligent Diagnosis System Of Urban Rail Transit Vehicle Based On Deep Learning

With the rapid development of urban rail transit in China,the contradiction between the increasing level of ridership development and the level of vehicle operation rate has become increasingly prominent.The level of vehicle operation rate largely depends on the level of improvement of vehicle operation and maintenance service.On the premise of the reliability of vehicle operation and maintenance,the research focuses on Intelligent Diagnosis System of urban rail transit vehicles,and it is of great significance to improve the level of urban rail transit vehicles service.Aiming at the characteristics of the current maintenance mode of urban rail transit vehicles,and the actual needs of fault diagnosis technology in the actual maintenance process,this research primarily constructs an Intelligent Diagnosis System for urban rail transit vehicles.Take importance into the aspects of vehicle intelligent diagnosis needs analysis,system composition,work flow and technical framework.The system takes Restricted Boltzmann Machine(RBM)as the basic structure of the deep neural network model,the fault feature vector of vehicle equipment as input vector,and the maintenance decision vector as output vector.The intelligent diagnosis model of urban rail transit vehicles is built based on Deep Belief Network(DBN).Contrast divergence algorithm is used to train the intelligent diagnosis model.Then,health assessment and fault warning for urban rail transit vehicles are provided.Taking the wheelset as an example,the intelligent diagnosis model of urban rail transit vehicles is trained and validated by its size data.The validation results show that the model has high diagnostic accuracy and short diagnostic time,which meets the requirements of accuracy and timeliness.The intelligent diagnosis system of urban rail transit vehicles in this research has value for application in the intelligent operation and maintenance of urban rail transit vehicles.