基于KNN的水电站水轮机监控系统研究  

Design of intelligent monitoring system for hydraulic turbine faults with improved K-proximity calculation and improved K-proximity algorithm

作　　者：谢科军 宋善坤 胡婷 姚娟 张利益 XIE Kejun;SONG Shankun;HU Ting;YAO Juan;ZHANg Liyi(Anhui Jiyuan Software Co.,Ltd.,Hefei 230061,China)

机构地区：[1]安徽继远软件有限公司,安徽合肥230061

出　　处：《粘接》2025年第1期193-196,共4页Adhesion

摘　　要：针对大型水轮机轴承故障诊断和预警准确率低,导致抽水蓄能电站存在状态监测与运维管理效果不佳的问题,提出一种大型水轮机轴承润滑油液在线监测系统。利用电涡流传感器对轴承油液数据采集,采用改进的K最近邻算法对轴承故障进行准确分类与诊断。结果表明,通过改进KNN算法,得到新故障与集合A中故障识别球的相似度最大值为0.4787,低于相似度匹配阀值0.6,说明改进KNN算法可实现新故障类型的准确识别,具备一定的自适应性和可扩展性;实际应用也进一步证明该算法可满足对水轮机轴承的状态监测、故障诊断和预警需求,实现水电站的准确监测和智能化运维管理。In response to the low accuracy of fault diagnosis and early warning of large water turbine bearings,which leads to poor monitoring and operation management of pumped storage power stations,a large water turbine bearing lubricating oil online monitoring system was designed.The eddy current sensor was used to collect the bear⁃ing oil data,and the improved K-nearest neighbor algorithm was used to accurately classify and diagnose the bear⁃ing faults.The results showed that the maximum similarity between the new fault and the fault identification ball in ensemble A was 0.4787,which was lower than the similarity matching threshold of 0.6,indicating that the im⁃proved KNN algorithm can achieve accurate identification of new fault types,and has certain adaptability and scal⁃ability.The practical application further proves that the algorithm can meet the requirements of condition monitor⁃ing,fault diagnosis and early warning of hydraulic turbine bearings,and realize accurate monitoring and intelligent operation and maintenance management of hydropower stations.

关 键 词：抽水蓄能电站 水轮机组 在线油液监测 K近邻算法 故障诊断 

分 类 号：TP392[自动化与计算机技术—计算机应用技术]