Research on rail transit DC switch trip fault diagnosis based on dual branch convolutional domain adversarial neural network

At present, the diagnosis of DC switch trip of rail transit is mostly based on traditional rules and experience, relying on manual rules or expert experience. However, due to the complex and dynamic rail transit system and diverse management needs, the traditional method is difficult to deal with the complex situation, resulting in low diagnostic accuracy. In order to optimize the accuracy of diagnosis, a fault diagnosis method of rail transit DC switch based on two-branch convolutional domain adversarial neural network is proposed in this paper. Firstly, the causes of DC switch trip are analyzed and the features are extracted. Secondly, according to the data collected and pre-processed under the 5 types of trip causes and normal operation conditions, an optimization algorithm based on the two-branch convolutional domain adversarial neural network is designed, and the algorithm is used to optimize the trip identification accuracy rate. Meanwhile, the advantages of the proposed new model are verified by comparing with various traditional fault diagnosis models. The experimental results show that the proposed model can effectively improve the accuracy of fault diagnosis.