基于聚类分析算法的发电异常数据监测与技术应用

打开文本图片集

中图分类号：TM835.4;TQ171.4 文献标志码：A 文章编号：1001-5922（2026）1-0270-04

Abstract：To further improve the precision and accuracy of power generation data monitoring and identification， this studyproposes touseclustering analysis to identifyabnormal power generationdataand employsoptical fiber sensors for monitoring.The experimental results show that with the increase of voltage，the voltage response time monitored by the optical fiber sensor gradually decreases.Uniformlycoating the optical fiber with anti-aging materials andadding equipotential ringscan improve the induction efciencyof voltage signals.When thevoltage is 35V，the response time of clustering analysis is only 1.0 s，which is much lower than that of the kNN algorithm，recursive algorithmand long short-term memory（LSTM）network algorithm.The sensitivityof the optical fiber sensor increases with the increase of voltage，with an increase amplitude of 98.18% . When the voltage is 35 V，the sensitivity is as high as 29.37MV/V.The monitoring precisionof abnormal power generation data based on the clustering analysis algorithm reaches 91. 33% ，and the average monitoring and identification accuracy exceeds 90% .Meanwhile，the variationrange of the misjudgment rateof power generation abnormal data monitoring and identification is small， with an average value of 9.8% ，which can meet the actual demand for abnormal data monitoring and identification.

Key words：cluster analysis;abnormal power generation data;monitoring and identification ;fiber optic sensor

对发电数据进行监测，及时发现异常数据，不仅可以提高电力生产的效率[1]，还能对发电企业的生产经营活动产生积极影响。（剩余5630字）