基于改进HHT的风力发电系统轴承故障诊断  

Fault Diagnosis for Rolling Bearing of Wind Turbine Based on Improved HHT

作　　者：沈艳霞[1] 王龙[1] 赵芝璞[1] 

机构地区：[1]江南大学电气自动化研究所,江苏无锡214122

出　　处：《测控技术》2013年第9期40-44,共5页Measurement & Control Technology

基　　金：国家自然科学基金资助项目(61104183);教育部新世纪优秀人才支持计划项目(NCET-10-0437)

摘　　要：轴承是风力发电机组中故障率较高的部件,其故障信号为非线性、非平稳信号,经验模态分解是一种自适应的信号处理方法,可用来分析此类信号,但是模态混叠使得经验模态分解无法准确地将固有模态分离出来。针对此问题,采用总体平均经验模态分解进行改进,利用高斯白噪声的频率均匀分布的统计特性,抑制模态混叠现象,并通过计算固有模态函数与故障信号的互信息来剔除虚假分量,从而得到更准确的Hilbert-Huang谱,由此提取故障信息,判断故障类型。仿真试验及轴承故障诊断实例均证明了该方法的有效性。Rolling bearing is an important component of wind turbine, which has a high failure rate. The fault signal of rolling bearing is nonlinear and non-stationary. Empirical mode decomposition（EMD） is an adaptive data analysis method which is able to process nonlinear and non-stationary signal. However, mode mixing makes it difficult to decompose the intrinsic mode functions accurately. Therefore, ensemble empirical mode decomposition（EEMD） is applied to solve this problem by utilizing the advantage of the statistical characteristics of white noise. The mutual information between IMFs and fault signal is calculated to distinguish the false modes. Therefore, a more accurate Hilbert-Huang spectrum is obtained, and the characteristics frequencies are extracted, then the fault is diagnosed. The simulation and experiments of real fault diagnosis verify the effectivehess of the proposed method.

关 键 词：风力发电系统 滚动轴承 故障诊断 Hilbert—Huang变换 总体平均经验模态分解 互信息 

分 类 号：TH165.3[机械工程—机械制造及自动化]