微分经验模态分解和能量矩阵相结合的水电机组故障信号预处理方法研究  

Research on the Fault Signal Preprocessing Method of Hydroelectric Unit by Combining Differential Empirical Mode Decomposition and Energy Matrix

作　　者：吴俊健 舒锦宏 徐灵江 吕延春 段文华 潘天航 WU Jun-jian;SHU Jin-hong;XU Ling-jiang;LÜYan-chun;DUAN Wen-hua;PAN Tian-hang(Jingshuitan Hydropower Plant,State Grid Zhejiang Electric Power Corporation,Lishui 320000,Zhejiang Province,China;Zhejiang Electric Power Dispatching Center,Hangzhou 310007,Zhejiang Province,China;NR Electric Co.,Ltd.,Nanjing 211102,China)

机构地区：[1]国网浙江省电力有限公司紧水滩水力发电厂,浙江丽水320000 [2]国网浙江电力调度控制中心,杭州310007 [3]南京南瑞继保电气有限公司,南京211102

出　　处：《中国农村水利水电》2022年第4期209-214,共6页China Rural Water and Hydropower

基　　金：国网浙江省电力有限公司科技项目(5211JS19001B)。

摘　　要：针对经验模态分解方法在处理水电机组振动这类由多重信号和噪声耦合的信号时出现的模态混叠现象,提出了一种微分经验模态分解和能量特征相结合的水电机组故障信号预处理方法。通过微分经验模态算法对信号进行分解,接着构造各阶微分信号能量矩阵,通过能量矩阵可以筛选出有效的分量。仿真试验表明微分经验模态分解能有效的避免模态混叠效应,使用支持向量机验证在水电机组导轴承信号预处理中,二阶经验模态分解构造的能量特征矩阵识别精度达98%。这一方法在水电机组复杂耦合信号预处理的应用中有很好的工程应用价值。In view of the modal aliasing phenomenon that occurs when EMD is used to deal with the vibration signals of hydropower units which are coupled by multiple signals and noise,a fault signal preprocessing method is proposed based on differential empirical mode decomposition and energy characteristics.The signal is decomposed by differential empirical mode algorithm and then the energy matrix of each order differential signal is constructed. The simulation results show that the differential empirical mode decomposition can effectively avoid the mode mixing effect. The support vector machine is used to verify that the identification accuracy can reach 98% in the preprocessing of the guide bearing signal of the hydropower unit. This method has a good engineering application value in the application of complex coupling signal pretreatment of hydroelectric units.

关 键 词：水电机组 微分经验模态分解 能量特征 信号预处理 

分 类 号：TV734.21[水利工程—水利水电工程]