摘要
应用包络分析于轴承故障诊断的关键是选择合适的滤波频带,特别是含有复杂结构的旋转机械设备,选择不同的滤波频带的包络分析结果相差很大。针对这一问题,提出基于复数小波多尺度分解和包络分析相结合的滚动轴承故障特征提取方法。利用复数小波分解实现信号的多尺度带通滤波,然后对小波系数求模,得到信号在各个尺度上的幅值包络,再对幅值包络求傅里叶变换,得到信号的二维幅值包络谱图,从而实现信号各个频段的快速包络分析,克服了经典包络分析方法需预知故障频带的缺点。对电站送风机实测振动数据分析表明,该方法具有突出的显示微弱故障信息的能力,可以有效地识别掩藏在强干扰噪声下的故障冲击成分;复小波多尺度包络谱图的横向切片与Hilbert包络谱具有非常相似的分析效果;而在旋转频率处的纵向切片对轴承故障最敏感,适用于滚动轴承微弱故障的报警。
The method of multi-scale enveloping spectrogram based on complex wavelet was proposed for feature value extraction of rolling bearing faults. It combines the band-pass filtering and envelope analysis of the measured vibration signals into a single-step operation and overcomes the shortcoming of the conventional envelope analysis in which the resonance frequency of the bearing should be known in advance. The case study of blower bearing shows that the proposed approach is suitable for identification of bearing fault under complicated running condition. The feature value from vertical slice of the multi-scale enveloping spectrogram at rotating frequency is the most sensitive to a bearing fault.
出处
《中国电机工程学报》
EI
CSCD
北大核心
2015年第16期4147-4152,共6页
Proceedings of the CSEE
基金
国家自然科学基金项目(51305135)
中国华能集团科技项目(HNKJ13-H20-05)~~
关键词
复数小波变换
多尺度包络分析
风机
滚动轴承
故障诊断
complex wavelet
multi-scale envelope analysis
blower
rolling bearing
fault diagnosis
