| 张 园，史永芳，迮素芳，李 力.基于邻域相关性小波去噪的滚动轴承包络解调及故障分类[J].轻工机械,2014,32(3): |
| 基于邻域相关性小波去噪的滚动轴承包络解调及故障分类 |
| Envelope Demodulation and Fault Classification of Rolling Bearing Based on Neighborhood Correlation Wavelet De-Noising |
| |
| DOI：IO. 3969/j. issn. 1005 -2895. 2014. 03. 004 |
| 中文关键词: 滚动轴承 邻域相关性 冗余第二代小波 包络解调 峭度 BP人工神经网络 |
| 英文关键词:rolling bearing neighborhood correlation redundant second generation wavelet envelope demodulation kurtosis BP artificial neural networks |
| 基金项目: |
|
| 摘要点击次数: 1433 |
| 全文下载次数: 1677 |
| 中文摘要: |
| 将邻域相关性的冗余第二代小波应用于滚动轴承信号降噪，用Hilhert包络解调法提取的故障特征频率，比较不
同转速和载荷下的提取效果，提出包络幅值峭度指标，并将其输入BP神经网络进行故障诊断。结果表明：基于邻域相关
性的冗余第二代小波降噪方法能很好的抑制噪声，保留原信号的信息；降噪后的故障信号经过Hilbert包络解调能找到
特征频率及其倍频，其效果优于原始信号的包络解调分析。工况会影响分析效果，且速度对提取效果的影响大于载荷。
包络幅值峭度指标能很好区分不同工况的故障信号，结合BP人工神经网络诊断正确率为100%。 |
| 英文摘要: |
| Rolling bearing vibration signals were de-noised based on neighborhood correlation and threshold value
redundant second generation wavelet. then rolling bearings' fault feature frequencies were extracted by Hilbert envelope
demodulation. and then feature extraction effect of different rotational speeds and loads were compared. Kurtosis of
envelope amplitude was proposed to input into BP artificial neural networks to make faults diagnosis. The result shows
that neighborhood correlation second generation wavelet de-noising method can well restrain the noise. and it can retain
the original signal information. The fault feature frequency and its frequency doubling can be basically found on the
Hilbert envelope demodulation of the de-noised signals. and the effect is prior to the original signal. Working condition
can influence the analysis effect. and the effect of velocity on extraction result is more obvious than load. The kurtosis of
envelope amplitude can well classify different working condition fault signals, and the diagnosis accuracy of kurtosis
index combined with BP artificial neural networks is 100% . |
| 查看全文 查看/发表评论 下载PDF阅读器 |
| 关闭 |
