An Early Warning Method of Distribution System Fault Risk Based on Data Mining

Table 2

Initial fault feature set of distribution system.

Category	Variable	Feature
Fault data	f1	Fault risk level
	f2	Monthly fault frequency
	f3	Number of households affected by power failure
Meteorological information	f4	Monthly cumulative rainfall
	f5	Monthly mean temperature
	f6	Monthly highest temperature
	f7	Monthly extreme weather days
	f8	Monthly mean humidity
	f9	Monthly extreme humidity days
	f10	Monthly gale days
	f11	Monthly thunderstorm days
	f12	Monthly snowy days
Operation data	f13	Maximum monthly load
	f14	Average monthly load
	f15	Month classification
Parameter data	f16	The geographical location classification
	f17	Feeder construction mode
	f18	Power supply area classification
	f19	Length of overhead feeder section
	f20	Length of cable feeder segment
	f21	Total length of the feeder
	f22	Number of feeder segment switches
	f23	Number of feeder transformers
	f24	Feeder operation time