Temporary faults in a medium-voltage distribution system may recur many times before they evolve into permanent faults. To accurately identify recurrent temporary faults can help to realize fault anticipation, reduce the occurrence of a permanent fault and enhance the power supply reliability. An equivalent circuit of a temporary fault is established and the temporary fault current waveform is analyzed. The results demonstrate that the zero-sequence current waveforms between recurrent faults are similar before they evolve into a permanent fault. A recurrent temporary fault identification method based on the zero-sequence current waveform similarity is proposed in this paper. An improved dynamic time warping (DTW) distance algorithm is used to calculate the general and transient waveform sequence distance of zero-sequence current. Then a waveform similarity is obtained by DTW distance to identify recurrent temporary faults. Amount of simulated data in PSCAD/EMTDC and field-measured data validate the accuracy and effectiveness of the proposed identification algorithm under varying influencing factors.
