Based On The S-transform Neural Network Automatic Blocking Line Failure Ranging

The automatic blocking and continuous railway power lines is a important part of electric system,it needs high circuit reliability. As the length of railway lines and ability of transport by train grow day by day,fast and accurate fault location of automatic blocking and continuous railway power lines play an important role to the safety and economic running of railway power system.the device of fault location of railway power system mostly follow low voltage network,these device mostly base on impedance measurement and others base on travelling wave, they can not aim at the peculiarity of automatic blocking and continuous railway power lines,so the precesion can not meet the needs of fault location of railway power system.A new approach of combined S-Transform and BP neural network are using for fault location of lOkV automatic blocking and continuous railway power lines in this paper.Firstly,the feature of automatic blocking and continuous railway power line's structure are analyzed,then,measuring signals by existing VT and TA of automatic blocking and continuous railway power lines, single phase to ground fault of automatic blocking and continuous railway power lines has been simulated in any different distances, any fault inception angle and variant resistances, The features of fault transient signals of voltage and current created by simulation extracted using S-Transform are put into BP neural network, then train and test the BP neural network,the result of trial is satisfied.Due to variant disturb signals contain a mass of inter-harmonics,its spectrum of frequency are not sure, so the amplitude of euploid harmonic is relatively small, making use of S-Transform to distill short window energy of euploid harmonic as inputs of neural network, eliminating variant disturb signals effectively.