| 武 鹏1 ，于明月 2.基于ITD和SVD的振动信号特征提取[J].航空发动机,2025,51(6):94-100 |
| 基于ITD和SVD的振动信号特征提取 |
| Vibration Signal Feature Extraction Based on ITD and SVD |
| |
| DOI： |
| 中文关键词: 本征时间尺度分解 Hankel 矩阵 奇异值分解 希尔伯特包络解调谱 故障诊断 |
| 英文关键词:intrinsic time-scale decomposition (ITD) Hankel matrix singular value decomposition(SVD) Hilbert envelope
demodulation spectrum fault diagnosis |
| 基金项目:航空动力基础研究项目资助 |
|
| 摘要点击次数: 2089 |
| 全文下载次数: 438 |
| 中文摘要: |
| 针对在含有大量气动噪声和结构噪声的机匣振动信号中提取故障特征的航空发动机轴承故障诊断的关键问题，提出一
种基于本征时间尺度分解（ITD）和奇异值分解（SVD）相结合的故障诊断方法。通过对在发动机外机匣测得的时域振动信号进行
ITD预处理，获取了对应的固有旋转（PR）分量，实现了振动信号中包含的故障特征的凸显；对各分量信号的时域信号通过系统延
迟法构建Hankel矩阵，并基于奇异值分解的奇异值差分谱方法进行降噪，以滤除噪声干扰，提取相对微弱的故障特征信息；选取
包含故障信息丰富的降噪后的PR分量用于信号重构，通过希尔伯特包络解调谱对重构信号进行特征提取，用于识别故障类型。
将此方法用于航空发动机轴承故障诊断，验证了其有效性。 |
| 英文摘要: |
| To address the critical challenge of extracting fault features from the casing vibration signals contaminated by significant
aerodynamic and structural noise in aeroengine bearing fault diagnosis, a fault diagnosis method integrating intrinsic time-scale decomposi?
tion (ITD) and singular value decomposition (SVD) is proposed. The method begins by applying ITD to preprocess time-domain vibration
signals measured from the outer casing, obtaining proper rotation (PR) components to highlight the fault features contained in the vibration
signals. A Hankel matrix is constructed using the system delay method for the time-domain signals of each component signal, and noise
reduction is performed based on the singular value difference spectrum calculated by SVD to filter noise interference and extract weak fault
characteristic information. The denoised PR components, which include rich fault information, are selected for signal reconstruction, and
fault feature extraction of reconstructed signals is carried out by Hilbert envelope demodulation spectrum for fault type identification. The
effectiveness of the proposed method was verified through its application to aeroengine bearing fault diagnosis. |
| 查看全文 HTML 查看/发表评论 下载PDF阅读器
|