基于经验模态分解和支持向量机的水电机组振动故障诊断.pdf

水力发电学报××××年第××卷第×期: ×- × Journal of Hydroelectric Engineering, ××××, Vol. ××, No. ×: ×- × 基于经验模态分解和支持向量机的水电机组振动故障诊断 李辉，李欣同，贾嵘，白亮，罗兴锜 （西安理工大学水利水电学院，陕西省西安市 710048） 摘 要：水电机组的振动信号为典型的非平稳、非线性信号。为了通过振动信号正确判断水电机组的运行状态， 本文提出运用经验模态分解处理原始信号，并对获得的基本模式分量计算其复杂度特征，最后运用最小二乘支 持向量机进行故障诊断。选取径向基函数作为核函数，并通过网格搜索和交叉验证确定相关参数。结果表明， 经验模态分解复杂度特征和支持向量机结合，能够准确的实现故障诊断，确定故障类型，为机组运行维护人员 提供参考依据。 关键词：水电机组；故障诊断；经验模态分解；复杂度；最小二乘支持向量机 中图分类号：TM312 文献标志码：A DOI: Fault Diagnosis of Vibration for Hydropower Unit Based on Empirical Mode Decomposition and Support Vector Machine LI Hui, LI Xintong, JIA Rong, BAI Liang,LUO Xingqi (Institute of Water ResourcesHydropower Engineering, Xian University of Technology, Xian 710048, Shaanxi Province, China) Abstract: The vibration signals of hydropower unit are typically non-linear and non-stationary. In order to judge accurately the running state of the hydropower unit through vibration signals, this paper proposes processing the original signals by applying empirical mode decomposition (EMD), and then calculating complexity features based on the obtained intrinsic mode functions. Finally, fault diagnosis is carried out by using the least squares support vector machine (LS-SVM). Radial basis function is selected as the kernel function, and relevant parameters are determined through grid search and cross validation. The results show that the combination of empirical mode decomposition (EMD) complexity features with support vector machine (SVM) can lead to accurate fault diagnosis, the determination of the fault type, and help provide a reference basis for the unit operation maintenance personnel. Key words