摘要
为提高水电机组故障诊断的准确性、确保机组安全运行,提出了基于局部切空间排列算法(local tangent space alignment,LTSA)与谱聚类的水电机组振动故障诊断方法。首先,利用GHM(Geronimo,Hardin,Massopust)多小波对机组振动信号进行分解,从所得多小波分解系数中提取多个特征参数,并利用检测指数(detection index,DI)对特征参数进行特征选择,降低特征维数;然后,利用LTSA对所得故障特征进行融合,获取低维强敏感特征参数;最后,将所得的特征参数输入到谱聚类算法中,实现水电机组振动故障识别。利用转子试验台和水电机组振动信号对所提出的方法进行了验证,结果表明该方法能够更好地识别机组故障。
To increase fault diagnosis accuracy of hydroelectric generating unit and guarantee its safe operation,a vibrant fault diagnosis method for hydroelectric generating unit based on local tangent space alignment(LTSA)and spectral clustering is proposed. Firstly, Geronimo, Hardin, Massopust(GHM) multi-wavelet is used to decompose vibrant signals of hydroelectric generating unit into several decomposed coefficients, from which multiple feature parameters are extracted and detection index(DI) is used to select the characteristic parameters to reduce the characteristic dimension.Then,LTSA is applied to fusion the selected feature parameters to obtain the low-dimensional and highly sensitive feature parameters.Finally, the obtained characteristic parameters are input into spectral clustering algorithm to realize the vibration fault identification of hydropower unit.In order to validate the effectiveness of the proposed method,signals acquired both from rotor test bench and hydroelectric generating unit are used to verify this method,and the results show that the proposed method can better identify faults of hydropower unit.
作者
卢娜
张广涛
刘付鑫
周廷鑫
边宏光
LU Na;ZHANG Guangtao;LIU Fuxin;ZHOU Tingxin;BIAN Hongguang(School of Hydro Science and Engineering,Zhengzhou University,Zhengzhou 450001,China;Rundian Energy Science and Technology Co.,Ltd.,Zhengzhou 450052,China)
出处
《武汉大学学报（工学版）》
CAS
CSCD
北大核心
2021年第11期1064-1069,共6页
Engineering Journal of Wuhan University
基金
国家自然科学基金项目(编号:51609203)。
关键词
水电机组
振动信号
故障诊断
局部切空间排列算法
谱聚类
hydroelectric generating unit
vibration signal
fault diagnosis
local tangent space alignment
spectral clustering
