Research On Voltage Level Identification Technology Of Transmission Line Based On Electric Field Detection

Electric shock is a major factor that threatens the safety o f living workers on transmission lines.In order to realize the electric shock prevention early warning of the high-voltage transmission line live operation,the pre-warning device on electric power field came into being at the historic moment.There are different critical electric fields near the safe distance because of the correspondence between the safety distance and voltage level.At present,most of these safety pre-warning devices Based on fixed threshold method,the automatic recognition of voltage level and intelligent warning function have not been achieved,so the utility of these devices cannot be guaranteed.This paper presents a voltage level identification method based on electric signal detection to improve the accuracy and practicability of safety pre-warning devices for these problems.First of all,this paper presents the current research status of voltage level identification and analyzes its advantages and disadvantages.After that,this paper presents a data fusion method of multi-electric field sensor to solve the problem that the result of one-dimensional electric field sensor is always affected by direction.Aiming at the problem that the low-precision decision tree has a negative impact on the classification result in the random forest model voting mechanism,this paper presents a voltage level identification method based on the weighted random forest.Then,this paper adds six new features,such as curve integration to improve the feature information on the basis of extracting polyno mial fitting coefficients for the problem of insufficient number of features.Finally,this paper combines the data fusion methods and weighted random forest algorithm proposed to realize the experimental evaluation of voltage level recognition.The result s show that the new features can significantly improve the recognition accuracy of 10 k V and 35 k V voltage levels.Using the weighted random forest algorithm proposed in this paper on the fusion electric field intensity value data,the recognition accuracy reaches94.07%,so the weighted random forest algorithm can provide an effective autonomous voltage level identification method for safety pre-warning devices.