汽轮发电机组振动故障诊断相关分析方法及应用  

Correlation Analysis Method and Application for Vibration Fault Diagnosisof Turbine Generator Unit

作　　者：孙和泰[1] 孙彬[1] 黄翔[1] 宋立远 杨建刚[2] SUN He-tai;SUN Bin;HUANG Xiang;SONG Li-yuan;YANG Jian-gang(Jiangsu Frontier Technology Co.Ltd.,Nanjing 211100,China;National Engineering Research Center of Turbo Generator Vibration,Southeast University,Nanjing 210096,China)

机构地区：[1]江苏方天电力技术有限公司,南京211100 [2]东南大学火电机组振动国家工程研究中心,南京210096

出　　处：《汽轮机技术》2020年第4期292-294,299,共4页Turbine Technology

基　　金：江苏方天电力技术有限公司“燃煤发电机组重要设备及重大危险源状态评估模型、事故诊断系统设计及应用(KJ201911)”资助。

摘　　要：从电厂SIS系统中提取数据,建立汽轮发电机组振动故障诊断相关和偏相关分析模型,分析变工况运行时各振动测点之间以及振动和过程参数之间的相关性规律,结合某600MW汽轮发电机振动故障实例进行了分析。研究结果表明,相关性特征反映了设备物理背景,可以有效地指导电厂设备故障诊断,提高数据挖掘的针对性。变工况运行时参数都在变化,参数之间的相互影响较大,偏相关分析更能展现出两两变量之间的净相关关系,可以减少故障诊断专项试验所需费用和时间。振动分析时的相关性特征包括:同一轴承座上轴振之间、轴振和瓦振之间、相邻轴承振动之间以及振动和过程参数之间等。如相关性质发生变化,需要进一步分析偏差的原因。The correlation and partial correlation analysis model for vibration fault diagnosis of turbo generator unit is established.Data is collected from SIS system of power plant to analyze the correlation between vibration and process parameters under variable operating conditions.An example of vibration fault of a 600MW turbo generator is presented.The results show that the correlation characteristics reflect the physical background of equipment.It can effectively diagnosis the fault and improve the pertinence of data mining.The parameters are changing simultaneously under variable operating conditions.The mutual influence between them is greater.Partial correlation analysis model can show the net correlation between the two variables better.It reduces the cost and time of special test for fault diagnosis.The correlative characteristics between the journal and the bearing vibration of the same bearing,the vibration between the adjacent bearings,and the relationship between vibration and process parameters,are used.If the correlation relationship changes,the reasons for the deviation need to be further analyzed.

关 键 词：汽轮机 发电机 振动 故障诊断 相关分析 

分 类 号：TM311[电气工程—电机]