Traffic safety has become a primary focus on the rapid development of railway traffic. Failure of railway turnout shall endanger train operations and affect their efficiency. Currently, the fault diagnosis of railway turnout still relies on the experience of maintenance personnel that can introduce several problems, such as low fault diagnosis efficiency and large amounts of required labor. To solve these problems, this paper takes the most widely used ZD6 turnout system in China as the research object, the railway turnout failure modes are summarized into eight typical types. Combined with the actual railway turnout fault data, the wavelet transform is used for the railway turnout action current extracted feature vector. The theory of fuzzy cognitive map (FCM) was designed to diagnose the fault modes of the railway turnout based on extracted feature vector, and the genetic algorithm particle swarm optimization (GAPSO) was selected to learn the weights of FCM. The simulation results indicated that FCM classifier model, based on GAPSO, could effectively classify the faults of the railway turnout. The GAPSO-FCM model could achieve a 97.1% fault diagnosis accuracy rate that showed that the GAPSO-FCM model proposed in this paper had high accuracy. Finally, we discuss further improvements that should be made to the model in the future.
Fault Diagnosis of ZD6 Turnout System Based on Wavelet Transform and GAPSO-FCM
2021-09-19
305644 byte
Aufsatz (Konferenz)
Elektronische Ressource
Englisch
Fault Diagnosis Method Based on Wavelet Transform
Springer Verlag | 2016
|Fault Diagnosis of Railway Turnout Based on Random Forests
Springer Verlag | 2020
|Fault Diagnosis of Railway Turnout Based on Random Forests
TIBKAT | 2020
|