摘要
为了解决轴承早期性能退化时信噪比低,特征提取和早期性能退化评估困难这一难题,本文采用盲源分离的方法分离轴承振动信号的干扰,将盲源分离后轴承振动信号的峭度值作为轴承性能评估的敏感特征,利用动态模糊神经网络建立轴承的早期性能退化模型。根据盲源分离后,早期性能退化时轴承振动信号的峭度值增加,可作为轴承早期性能退化评估的敏感特征。计算结果表明,盲源分离使得振动信号的峭度对轴承的性能状态更加敏感,轴承性能退化评估结果准确,具有重要的工业实用价值。
Signal-to-noise rates( SNR) is very poor when incipient faults occur on rolling bearings,which makes it difficult to extract fault features and evaluate performance degradation of rolling bearings. Blind source separation( BSS) method is adapted to separate noise impulses mixed in measured signals of rolling bearings. Kurtosis values are sensitive features when measured signals are separated by BSS method. A performance evaluating model of incipient degradation of rolling bearings is built based on dynamic fuzzy neutral network( DFNN) and kurtosis values are input vectors of the evaluating model. The results of calculation show that the kurtosis is more sensitive to the performance and evaluation results of degradation performance of rolling bearings are more accurate. The method of performance degradation based on kurtosis and BSS is significant to solve engineering problems.
出处
《机械科学与技术》
CSCD
北大核心
2017年第11期1771-1777,共7页
Mechanical Science and Technology for Aerospace Engineering
基金
国家自然科学基金项目(51275426)
四川省科技计划项目(2015HH0015)
中央高校基本科研业务费专项资金资助(2682014CX034
2682014BR024)资助
关键词
滚动轴承
早期性能退化
盲源分离
性能评估
峭度
rolling bearing
incipient performance degradation
blind source separation
performance evaluation
kurtosis