Research on switch fault diagnosis based on LTTB algorithm and LSTM model
LI Li
Abstract：With the increase in the speed and density of rail transit trains,the failure rate of railway switches is on the rise,severely impacting operational efficiency.To promptly detect switch failures,accurately diagnose fault types,and thereby improve maintenance quality and reduce costs,this paper proposes a fault diagnosis method based on the LTTB algorithm and the LSTM model.The paper analyzes the characteristics of switch action curves and the causes of faults,employs the LTTB down-sampling algorithm and an interpolation method to integrate the data dimensions of the action curves,and performs model training and analysis using the LSTM algorithm.The results show that this method can quickly and accurately identify switch fault types based on their action curves,thereby enhancing the work efficiency of maintenance personnel.
Keywords：LSTMLTTBswitchfault diagnosisdata dimension integrationcurve of action
Publication Date：2025-11-25
Pages：5( 53-57 )