The invention discloses a grounding fault start method based on
power grid characteristic parameter continuous increment, and the method comprises the steps: detecting the displacement
voltage U<N> of a neutral point in a distribution network and the zero sequence current I<0N> of each
feed line in real time; determining that a
power grid has a single-phase grounding fault if, in a time period
delta T, the stable increment of the
voltage of the neutral point in the distribution network accords with a formula:
delta U<N> >=
delta U<NL>, and the stable value of the zero sequence current of any one
feed line accords with a formula: I<0N> >= I<0E>, and enabling a grounding fault
processing device to carry out the fault start at this moment, and to process the single-phase grounding fault; determining that the changes of
power grid parameters cause the change of the displacement
voltage U<N> of the neutral point if, in a time period delta T<1>, the stable increment of the voltage of the neutral point in the distribution network accords with the formula: delta U<N> >= delta U<NL>, but the maximum zero sequence current of the feed lines accords with a formula: I<0N> <= I<0E>, and enabling the grounding fault
processing device not to start; and determining that the fault is an instable disturbance grounding fault or an intermittent grounding leakage fault if, in a time period delta T, the stable increment delta U<N> of the voltage of the neutral point in the distribution network does not stably accord with the formula: delta U<N> >= delta U<NL>, or the maximum zero sequence current I<0N> of the feed lines does not stably accord with the formula: I<0N> <= I<0E>.