State Identification And Fault Early Warning Method For Live Working Of Distribution Line Based On Data Analysis

In view of the problems such as frequent changes in data status during live operation of distribution lines,lack of real-time monitoring of lines,and difficulties for live operation personnel to accurately identify early warning due to equipment faults during live operation,this thesis analyzes the data of distribution lines,conducts early warning research on the state of live operation of distribution lines,and proposes the method of line status recognition and live operation fault early warning.Reduce the impact of distribution line faults on the life of power users,improve the safety of actual operation of distribution line live operation,and ensure the normal power consumption of users.In this thesis,relevant sensors are used to collect all kinds of running state data of distribution lines,and the data transmission process is optimized to reduce the data transmission time.For the problem of mixing abnormal data collected in the running state of distribution lines with abnormal data of lines,Thompson-tau method combined with quartile algorithm was used to identify abnormal values in the running state of distribution lines and optimize the measured data of live operation of distribution lines.The running status data of distribution lines are analyzed in depth,and the factors that have great influence on distribution lines are compared first to obtain the low-dimensional and high-precision sample data set of live operation of distribution lines.Secondly,to solve the problem that distribution line equipment faults and hidden dangers are difficult to be accurately identified and warned,a distribution line data prediction method based on sample and variable double weighting and abnormal data identification method considering correlation coefficient are proposed to judge the abnormal phenomena of the line.Finally,a condition warning method based on live operation of distribution lines is established.This method uses naive Bayes algorithm to mine the data of distribution lines and live operation,and establishes the fault factor database of distribution lines in live operation.Through the method of time series similarity fault matching,a fault discriminant database for live operation of distribution lines is established,fault factors are compared,At the same time,the intelligent auxiliary system for live working of distribution line is developed to realize the safety early warning of live working data of distribution line,the operators carry out safety operation according to the early warning,improve the safety of live-line operation,and meet the timeliness and accuracy of fault early warning.The innovation of this thesis is that through the data analysis of the live operation of the whole distribution line,the status identification of the distribution line and the fault warning of live operation are completed,so as to ensure the normal electricity consumption of users,and make up for the deficiencies of the current live operation monitoring means of the distribution line,the long time of manual monitoring and the high cost of machine monitoring.