电力分析中的异常报警数据挖掘技术优化

打开文本图片集

中图分类号：TP277；TM711 文献标志码：A 文章编号：1001-5922（2025）09-0177-03

Optimization of abnormal alarm data mining technology in power analysis

SHU Chang（Guangdong Power Grid Customer Service Center，Guangzhou 671OOO，China）

Abstract：Inorder to further improve the eficiencyof abnormal alarmdata mining，the clusteranalysis technology was proposed to mine the abnormal alarm data in the power data to improve the accuracyof power analysis，and the parameters such as data mining time，eror rate and delay were used to further verifythe effect of clustering analysis in abnormal alarm data mining.Simulation results showed that the recall rateof clustering algorithm was always higher than thatof convolutional neural network algorithmand deep learning algorithm，with themaximumrecalrate of 99.4% and the minimum recall rate of 98.7% ，and the average recall rate of clustering algorithm was still as high as 99.04% .The maximum data mining time of clustering analysis technology was 1.2 s，while the maximum data mining time of convolutional neural network algorithm and deep learning algorithmwas1.9and 2.2 s，respectively，which was 36.84% and 45.45% higher than that of clustering analysis technology，respectively.It can meet the requirements of actual abnormal alarm data mining.

Key Words：cluster analysis;technical optimization;abnormal alarm data;excavation time;mining error rate

近年来，电力数据库中积累了数以亿计的数据信息，存在许多有价值的信息[1]。（剩余4533字）