Fault Diagnosis Method of Track Circuit Based on PSO-LSTM

Track circuit system is widely used in the field of railway transportation, due to the large system and the wide variety of faults, and the current fault diagnosis of track circuit mostly considers the spatial information of the fault data, but not its time information, and some fault types are related to the time series data, so this paper proposes a PSO-LSTM model based on nonlinear weight improvement to process the fault data in time series information and diagnose a variety of fault types of the track circuit. The improved PSO optimization algorithm is used to find the optimal initial value suitable for the parameters of the LSTM model, so as to effectively improve the fault diagnosis ability of the model. Finally, through the training and testing experiments of the measured data, it is shown that the improved model has a more significant fault diagnosis effect than the original model, that is, the improved model is more effective in the fault diagnosis of track circuits.