Remote short-circuit current determination in DC railway systems using wavelet transform

For effective protection of DC railway systems, it is important to discriminate the remote short-circuit current from the train starting current. Four methods for the discrimination are assessed. The first two methods apply the traditional heuristics by comparing the magnitude and the rate of change (diidt) between the two currents. The remaining methods use the Fourier and wavelet transforms. It is shown that the first three methods do not provide consistent discrimination. The remote short-circuit current is determined mainly by the steel rail impedance, which is time varying due to the skin effect. In contrast, impedances of the traction motor and the contact wires, and the change of operating mode during starting govern the train starting current. The wavelet transform identifies these salient features. The remote short-circuit current is simulated using EMTP, and a SIMULlNK model is used to obtain the train starting current.