摘要
希尔伯特-黄变换能精确描述信号频率和幅值随时间的变化,高阶累积量可以自动地抑制高斯背景噪声。为正确分析机械故障领域所采集的信号,将高阶谱和HHT的优点有机结合,通过EMD分解得出三阶累积量的各个本征模态函数分量,通过Hilbert变换得出相应的时频特性曲线和幅频特性曲线,并根据幅频特性曲线进行故障诊断,然后将诊断结果和原始信号进行比较。比较结果表明,基于高阶累积量和HHT的机械故障诊断方法是可行的和必要的。
The Hilbert-Huang transform can be used to describe the signal frequency and amplitude changes in time domain pre- cisely. The high order cumulant can be used to reduce the Gaussian background noise automatically. In order to analyze the signals of mechanical faults, the advantages of high order spectrum and HHT were combined effectively, intrinsic mode functional components of third-order cumulant were obtained by EMD decomposition, and time-frequency and amplitude-frequency curves were drawn by Hilbert transition. The fault analysis was carried out based on the obtained amplitude-frequency curves. By comparing the experimental results with primitive signals, it proves that the experimental method is feasible and necessary.
出处
《机床与液压》
北大核心
2012年第23期175-177,188,共4页
Machine Tool & Hydraulics
基金
国家自然科学基金资助项目(50975098)
