摘要
本文基于经验小波变换(EWT,empirical wavelet transform)和奇异值分解(SVD,singular value decomposition)技术提出了一种齿轮的故障诊断方法.首先采用EWT方法将齿轮的振动信号分解为若干个本征模态分量(IMF),并利用这些IMF分量形成向量矩阵.而后对初始向量矩阵进行奇异值分解,根据奇异值分解的三大特性,将求得的特征向量矩阵的奇异值作为齿轮振动信号的模式特征向量.最后通过建立马氏距离判别函数判断齿轮的振动情况和故障类型.通过对实际实验数据的分析,证明了该方法在齿轮故障诊断中有效性.
Based on the empirical wavelet transform（EWT） and the singular value decomposition（SVD） tech- nique, a gear fault diagnosis method is proposed. The EWT method is used to decompose the vibration signal of the gear into several intrinsic mode function（IMF） signal, which are used to form the feature vector ma- trix. Then the singular values of the original vector matrix is obtained by SVD, which is taken as the pattern feature vector of the gear vibration signal. Finally, the vibration condition and the fault type of the gear are judged by establishing the Markov distance discriminant function. Through analyzing the actual experimental signals, it is proved that the method is effective in gear fault diagnosis.
出处
《三峡大学学报（自然科学版）》
CAS
北大核心
2018年第1期80-85,共6页
Journal of China Three Gorges University：Natural Sciences
基金
国家自然科学基金项目(51205230)
湖北省重点实验室开放基金课题(2016KSD15
2016KSD14)
三峡大学人才科研启动基金项目(KJ2012B014)
关键词
EWT
SVD
齿轮
故障诊断
特征向量
empirical wavelet transform （EWT）
singular value decomposition （SVD）
gear
fault diagnosis
feature vector
