基于EEMD-KECA的风电机组滚动轴承故障诊断

发布时间：2018-01-09 20:11

  本文关键词：基于EEMD-KECA的风电机组滚动轴承故障诊断　出处：《太阳能学报》2017年07期 　论文类型：期刊论文

  更多相关文章： 故障诊断 聚合经验模态分解 核熵成分分析 能量熵 滚动轴承

【摘要】：针对传统频域诊断算法不能充分挖掘出非线性、非平稳信号内部本质信息的问题,提出基于聚合经验模态分解(EEMD)的复合特征提取和基于核熵成分分析(KECA)的故障自动诊断算法。该方法首先采用EEMD将原始信号分解成若干特征模态函数(IMF),计算IMF能量和信号的能量熵构建复合特征向量并作为KECA的输入,之后建立KECA非线性分类器并引入一种新的监测统计量——散度测度统计量,实现故障的实时监测与自动诊断。采用KECA可实现根据熵值大小进行特征分类,具有较强的非线性处理能力,且不同特征信息之间呈现出显著的角度差异,易于分类。最后通过实际风电机组滚动轴承应用实例对算法进行验证,结果表明该方法可有效提取信号中的故障特征,实现对滚动轴承的故障诊断,相比神经网络分类方法具有更高的识别率。
[Abstract]:To solve the problem that the traditional frequency domain diagnosis algorithm can not fully excavate the internal essential information of nonlinear and non-stationary signals. Composite feature extraction based on polymeric empirical mode decomposition (EEMD) and kernel entropy component analysis (KECA) are proposed. The method firstly decomposes the original signal into several characteristic mode functions by EEMD. The energy entropy of IMF and signal is calculated to construct the compound eigenvector as the input of KECA, and then the nonlinear classifier of KECA is established and a new statistical measure of divergence is introduced. KECA can be used to classify features according to entropy value, which has strong nonlinear processing ability, and there are significant angle differences among different feature information. It is easy to classify. Finally, the algorithm is verified by the actual wind turbine rolling bearing application example. The results show that the method can effectively extract the fault features from the signal and realize the fault diagnosis of the rolling bearing. Compared with the neural network classification method, it has a higher recognition rate.
【作者单位】： 内蒙古工业大学电力学院;内蒙古北方龙源风力发电有限责任公司;北京工业大学电子信息与控制工程学院;
【基金】：国家自然科学基金(61364009;21466026) 内蒙古自然科学基金(2015MS0615) 校级重点项目(X201424)
【分类号】：TH133.3;TM315
【正文快照】： 0引言滚动轴承是旋转机械的主要部件之一,具有效率高、摩擦阻力小、装配简单、易润滑等优点,因此被广泛应用于风力发电机传动链系统,是该系统中应用最普遍、使用最多,也是最易损伤的部件之一。风电机组传动链中的许多故障都与滚动轴承有密切关系,据统计约30%的机械故障与轴承，

本文编号：1402460