基于迭代的集总经验模式分解算法的齿轮箱故障特征提取

Fault Feature Extraction of Gearbox based on Iterative Ensemble Empirical Mode Decomposition Method

针对集总经验模式分解方法(Ensemble Empirical Mode Decomposition,EEMD)在实际应用中存在的盲目添加白噪声的问题,提出了一种迭代的集总经验模式分解方法(Iterative Ensemble EmpiricalMode Decomposition,IEEMD)。首先介绍了IEEMD方法,然后将EEMD方法与IEEMD方法用于实际齿轮箱故障信号的特征提取。结果表明,与EEMD方法相比,IEEMD方法可以得到更高分辨率的HHT时频谱图,可以从信号中提取更多的有用信息。说明IEEMD方法较好地克服了EEMD方法中存在的盲目添加白噪声的问题,适合于作为齿轮箱故障信号的特征提取方法。

Aimed at the drawback of blindly adding white noise occurring in the ensemble empirical mode decomposition（EEMD） when used to solver realistic problems,the iterative ensemble empirical mode decomposition（IEEMD） method is proposed.To begin with,the IEEMD method is introduced.Then,the EEMD method and the IEEMD method are used to analyze the signals from the realistic gearbox with a broken-tooth fault.As a result,the comparisons with the EEMD method show that the IEEMD method could produce HHT spectrum with higher time-frequency resolution and extract more and useful information from the signals.Moreover,it indicates that the IEEMD method greatly alleviates the drawback of the EEMD method and is suitable as a fault feature extraction method for gearboxes.