The contact network undetakes the important task of delivering electric energy to the electric locomotives along the line and

once in case of fault

will directly affect safe operation of the train

so it is necessary to study the fault identification method of the contact network. In this paper

a fault identification method based on VAE and TabNet is propsoed. Firstly

ANSYS Workbench software is used to set up the pantograph - network coupling model to obtain the variation of pantograph - network contact force under different faults; Then

the variational self - encoder algorithm is used to expand the fault simulation data and improve the generalization ability of the model. Finally

the identification of fault types is performed by TabNet model with an identification accuracy rate of 96%. It is also compared with other traditional classification algorithms. The experimental results show that the proposed method exhibits superior performance in fault identification of the contact network.