中文摘要：

水轮机轴系信号是一种典型的时变非平稳信号,常规的Hilbert-Huang变换方法虽能良好地分析这类信号,但原始信号经EMD分解后,在低频区易产生虚假的IMF分量,直接影响特征信息的提取,导致误判。本文提出基于能量相关波动的改进HHT方法及其判别条件。通过仿真信号对该方法进行了有效性验证,并以原型水轮机非最优工况下动态信号为例,进行了应用检验。结果表明,与常规方法相比,该方法便捷有效、并具有良好的虚假分量识别能力,提取真实的水轮机特征参量,更加适合分析复杂而特殊的水轮机动态特征信息。

英文摘要：

Hydroturbine shafting signal is a typical time-varying non-stationary signal which can be analyzed by conventional HHT.After decomposing the original signal by the EMD,however,in the low frequency zone it is easy to generate an artificial IMF component which would directly impact the feature extraction and lead to a wrong judgment.In this paper an improved HHT method and its discrimination are put forth on the basis of energy-correlation fluctuation.This method is validated by using simulation signals of a dynamic signal test for prototype hydroturbine under non-optimal operating condition.The results show that the new method is more convenient and effective than conventional HHT and can well identify false component and extract true characteristic parameters.The method is useful in analyzing the complex and specific dynamic characteristics of hydroturbine.