摘要
结合经验模式分解(EMD)、随机减量技术(RDT)和遗传算法(GA),提出一种用于旋转机械油膜稳定性在线监测的特征提取方法。用经验模式分解从原始振动信号中分解出与油膜稳定性有关的低频分量,用随机减量技术从该低频分量中提取出自由衰减响应,然后采用遗传算法从自由衰减响应中拟合出系统的阻尼比作为稳定性参数。方法已应用于离心压缩机组油膜稳定性的在线监测,结果表明用本方法提取的特征参数能够有效地反映油膜稳定性状态,可对油膜稳定性裕度进行趋势分析,实现稳定性的在线监测和预报。
Based on empirical mode decomposition (EMD), random decrement technique (RDT) and genetic algorithm (GA), a method is presented for extracting oil-film stability parameters from vibration signals in rotating machinery, which is used for on-line monitoring and analysis of oil-film whirl problem. By using EMD, the low-frequency component is firstly decomposed from original signal, and then free decay response is extracted from the low-frequency signal through RDT. As a feature parameter, damping ratio is obtained from the free decay response using GA. This method is now used for the on-line monitoring of the stability problem for oil-film whirl of a carbon dioxide compressor. The result shows that the feature parameter extracted through the proposed method can effectively represent the condition of oil-film stability, and thus can be used for the on-line prediction of the onset of oil-film whirl.
出处
《振动工程学报》
EI
CSCD
北大核心
2006年第4期446-451,共6页
Journal of Vibration Engineering
基金
国家自然科学基金资助项目(50475087)
关键词
旋转机械
在线监测
油膜涡动
经验模式分解
随机减量技术
rotating machinery
on-line monitoring
oil-film whirl
empirical mode decomposition
random decrement technique