全景动态网络标志物的汽轮发电机定子绕组热故障预警  

Early Warning of Thermal Fault in Turbine Generator Stator Winding of Landscape Dynamic Network Marker

作　　者：曾思嘉 方瑞明[1] 彭长青[1] 庄杰农 尚荣艳[1] ZENG Sijia;FANG Ruiming;PENG Changqing;ZHUANG Jienong;SHANG Rongyan(College of Information Science and Engineering,Huaqiao University,Xiamen 361021,China)

机构地区：[1]华侨大学信息科学与工程学院,福建厦门361021

出　　处：《华侨大学学报(自然科学版)》2025年第2期201-208,共8页Journal of Huaqiao University(Natural Science)

基　　金：国家自然科学基金资助项目(52477048);福建省高校产学合作项目(2024H6009);福建省厦门市自然科学基金资助项目(3502Z202373952);福建省厦门市产学研项目(2023CXY0201)。

摘　　要：将汽轮发电机组的集散控制系统(DCS)的定子各槽出水口水温监测点映射为复杂网络中的节点,从而能够基于汽轮发电机DCS监测数据对定子绕组的热状态进行观测。根据DCS监测数据的时序特性,引入全景动态网络标志物(L-DNM)法计算网络中各节点的特异性皮尔逊相关系数,以构建不同采样时刻的特异性差分网络。量化网络中各节点的动态变化以进行故障预警,进而筛选出温度异常升高的关键节点,根据这些关键节点构建动态网络标志物(DNM)以识别故障位置。结果表明:文中方法能够实现对早期故障的预警和异常槽口位置的定位。The outlet water temperature monitoring points for each stator slot in the steam turbine generator group distributed control system(DCS)are mapped to nodes in a complex network,the thermal state of the stator windings can be observed based on the steam turbine generator DCS monitoring data.Based on the time series characteristics of the DCS monitoring data,the landscape dynamic network marker(L DNM)method is introduced to calculate the specific Pearson correlation coefficients of each node in the network to construct specific differential networks at different sampling times.The dynamic changes of each node in the network are quantified for the purpose of fault prediction.Subsequently,critical nodes with abnormal increasing temperature are identified to construct a dynamical network marker(DNM)for fault location identification.The results show that the proposed method can achieve early warning of faults and localization of abnormal slot positions.

关 键 词：汽轮发电机 定子绕组 热故障 全景动态网络标志物 故障检测 故障定位 

分 类 号：TM311[电气工程—电机]