ELECTRICA

THE MODELLING RECORDED FAULTS IN RAILWAYS AND PREDICTION

1.

Turksat Satellite Communication and Cable TV Operation Joint Stock Company, Golbasi, Ankara, TURKEY

2.

Istanbul University Faculty of Engineering 34320, Avcilar, Istanbul, TURKEY

ELECTRICA 2017; 17: 3417-3423
Read: 871 Downloads: 569 Published: 27 July 2017

This paper describes theprediction of the accidents and incidents by modelling the faults that are hardto be predicted, and forms the stochastic processes in railways. As it is seen,the root causes of the occurrence of the accidents and incidents in railwaysare the technical faults, the organizational faults, the staff faults, thethird-party faults, the other faults and some parameters affecting these faultsholidays, seasonal effects etc. The system model was set up by considering theabove-mentioned faults and the causes as the inputs of the model and theaccidents and the incidents as the outputs of the model. The real values of theaccidents and the incidents were compared with the predicted values by usingArtificial Neural Networks-ANN and Support Vector Machines-SVM. The results,which were acquired by this article, are considered as a guide because of theabsence of any studies about the prediction of the accidents caused by thefaults.

Files
EISSN 2619-9831