摘要
针对风电机组齿轮箱中齿轮故障特征提取与故障诊断问题,提出一种基于集合经验模式分解(EEMD)、奇异谱熵和模糊C均值聚类的故障诊断方法。首先对原始振动信号进行EEMD分解,得到各阶本征模态函数(IMF)构成的特征模式矩阵。接着对该特征模式矩阵求奇异谱熵值,奇异谱熵值的大小能反映部件的工作状态和故障类型。最后,将得到的奇异谱熵值矩阵进行模糊聚类分析并得到分类结果。通过对齿面磨损、齿面剥落和正常3种齿轮状态分别使用EMD法和EEMD法进行故障分类对比,结果验证了该方法的有效性和可行性,同时证明EEMD法具有更好的分类效果。
In order to solve the problems of fault feature extraction and fault diagnosis of gears in wind turbine' s gearbox, the fault diagnosis method using ensemble empirical mode decomposition (EEMD) , singular value spectral entropy and fuzzy C means cluster (FCM) was proposed. Firstly, the EEMD method was used to decompose the original vibration signal of gearbox so that an initial feature vector matrix with intrinsic mode functions was formed. Then the singular spectrum entropy of the feature pattern matrix was calculated, as the size of singular spectrum entropy can reflect the system' s working condition and fault type. Finally, the FCM cluster method was used to analyze the singular value spectrum entropy matrix and get the result of final fault classification. The EMD method and the EEMD method were used to compare fault classification of three gear statuses, including gear surface abrasion, gear surface chipped and normal, respectively. The results prove the validity of proposed method, furthermore, the EEMD method is more effectively.
出处
《太阳能学报》
EI
CAS
CSCD
北大核心
2015年第2期319-324,共6页
Acta Energiae Solaris Sinica
基金
国家自然科学基金(51279161)
关键词
风电机组
齿轮箱
故障诊断
集合经验模式分解
奇异谱熵
模糊C均值聚类
wind turbines
gearbox
fault diagnosis
ensemble empirical mode decomposition
singular value spectral entropy
fuzzy C means clustering