The application relates to the technical field of
power system protection, and specifically discloses a single-phase grounding fault line
selection method for a distribution network based on
deep learning, which comprises the following steps: firstly, according to the influence of the fault condition on the transient zero-mode current at the outlet of a medium-
voltage feeder, a field fault test or a
simulation scheme is designed; secondly, the transient zero-mode current data are recorded through tests or
simulation of multiple medium-
voltage feeders; thirdly, a three-dimensional
time sequence matrix is obtained through preprocessing, and a
training set and a
test set are divided and normalized; fourthly, a global-single-channel
hybrid model is constructed based on 1DCNN,
time sequence features among multiple channels and in a single channel are extracted, and the
time sequence features are fused through a fully-connected neural network; and finally, the model is trained in combination with a data enhancement strategy, and the
test set is verified to realize fault line selection. The single-phase grounding fault line
selection method for the distribution network based on
deep learning can significantly improve the line selection accuracy and reliability, the actual test or
simulation scheme is used to guarantee the
data quality, the model can accurately extract features, the fault can be quickly isolated, and the
safe operation of the distribution network is guaranteed.