基于SVD优化LMD的电梯导靴振动信号故障特征提取

发布时间：2018-03-18 08:36

本文选题：电梯导靴　切入点：奇异值分解　出处：《振动与冲击》2017年22期 　论文类型：期刊论文

【摘要】：针对局部均值分解(Local Mean Decomposition,LMD)方法提取电梯导靴振动信号的故障特征分量时存在的模态混淆现象,本文提出了一种基于奇异值分解(Singular Value Decomposition,SVD)优化局部均值分解(Local Mean Decomposition,LMD)的电梯导靴振动信号故障特征提取方法。该方法以奇异值贡献率原则构造原始信号的Hankel矩阵,采用SVD对Hankel矩阵进行分解;将曲率谱原则与奇异值贡献率原则相结合对奇异值进行选择,将包含主要故障信息的奇异值进行逆重构,得到剔除噪声信号与光滑信号的突变信号;并利用LMD方法对突变信号进行故障特征提取,得到能够突出原始信号振动特征的故障特征分量。实例结果表明该方法有效改善了LMD的模态混淆现象,更准确地提取了振动信号的故障特征分量,为电梯导靴的故障诊断提供了一条有效的途径。
[Abstract]:For the local mean decomposition (LMD) method to extract the fault characteristic components of the vibration signals of elevator guide boots, there is modal confusion phenomenon in this paper. In this paper, a method based on singular Value decomposition (SVD) to optimize local mean decomposition (LMD) for vibration signal fault feature extraction of elevator guide shoe is presented. The Hankel matrix of original signal is constructed by using singular value contribution rate principle. The Hankel matrix is decomposed by SVD, the singular value is selected by combining the curvature spectrum principle with the singular value contribution principle, and the singular value which contains the main fault information is inversely reconstructed, and the abrupt signal which removes the noise signal and the smooth signal is obtained. The fault feature components which can highlight the vibration characteristics of the original signal are obtained by using the LMD method to extract the fault feature of the sudden change signal. The results show that the method can effectively improve the modal confusion phenomenon of the LMD. The fault characteristic components of vibration signals are extracted more accurately, which provides an effective way for the fault diagnosis of elevator guide boots.
【作者单位】：昆明理工大学质量发展研究院;云南省特种设备安全检测研究院;昆明理工大学信息工程与自动化学院;
【基金】：国家质检总局科技计划项目资助(2013QK104) 云南省质量技术监督局科技计划项目资助(2013YNZJKJ02)
【分类号】：TU857

【相似文献】