Electricity stealing identification method based on a support vector machine

Abstract

The invention discloses an electricity larceny identification method based on a support vector machine, and relates to an electricity larceny identification method. The current electricity larceny prevention technology only depends on monthly electric quantity change of a user and line loss abnormity of a transformer area to determine suspected electricity larceny households, and has the problemsof poor timeliness, low accuracy, and low efficiency when electricity larceny behaviors are found through on-site investigation. The method comprises the steps of fluctuation ratio calculation, normalload data sample selection, establishment of an SVM normal load data classification model according to the normal load data samples, electricity stealing identification according to the SVM normal load data classification model, separation of suspected outliers of electricity stealing, determination of electricity stealing sample points and setting of electricity stealing point alarms. Accordingto the technical scheme, the novel electricity stealing identification method is combined with a method for calculating the electric quantity fluctuation rate and a support vector machine analysis method. Appropriate sample data are selected for the support vector machine by calculating the electric quantity fluctuation rate, so that adverse effects of sample problems on a detection analysis result can be effectively reduced, and an electricity stealing detection result is more accurate.

Description

technical field [0001] The invention relates to a method for identifying electricity theft, in particular to an identification method for electricity theft based on a support vector machine. Background technique [0002] With the emergence of high-tech means of stealing electricity, the problem of stealing electricity has become more and more prominent. The annual economic losses caused by electricity theft in the country are about tens of billions of yuan, which has seriously endangered the normal operation of the economic order. The current anti-stealing technology only relies on the monthly power changes of users and the abnormal line loss in the station area to determine the suspected electricity theft, which has the problems of poor timeliness, low accuracy, and low efficiency in on-site investigations to find electricity theft. Contents of the invention [0003] The technical problem to be solved and the technical task proposed by the present invention are to perfect...

AI Technical Summary

Problems solved by technology

The current anti-stealing technology only relies on the monthly power changes of users and the abnormal line loss in the station area to determine the suspected users of electricity theft, which has poor timeliness and low accuracy, and the efficiency of on-site investigations to find electricity theft is not high.

Images