摘要
针对管道阀门内漏声发射信号的非平稳特征,提出将总体平均经验模式分解(EEMD)和希尔伯特变换相结合的希尔伯特黄变换(HHT)方法应用于管道阀门内漏声发射检测中。利用EEMD方法将阀门内漏声发射信号自适应分解为一簇本征模态函数(IMF),并对分解后的信号进行Hilbert谱和HHT边际谱分析,可以提取到阀门内漏声发射信号的本质特征,突破常规时频分析的非线性信号局限性以及经验模式分解(EMD)造成的模态混叠现象;相比于STFT频谱,Hilbert谱具有较高的时频分辨率,最后对信号的Hilbert边际谱分析确定了内漏信号的主要作用频率。该方法可以有效地应用于阀门内漏声发射检测中。
To analyze the non-stationary feature of acoustic emission （AE） signal from pipeline valve leakage, a HHT method combining ensemble empirical mode decomposition （EEMD） and Hilbert transformation method was presented here. Using EEMD method, the detected leakage signal was self-adaptively decomposed into a series of intrinsic mode functions（IMFs）. Then, those IMFs were analyzed using Hilbert spectrum and HHT marginal spectrum, the key characteristics of AE signal in valve leakage were extracted. It was shown that the proposed method can overcome limitations of nonlinear signal in time-frequency analysis, and can avoid the mode mixing caused by EMD method; the Hilbert spectrum has higher time-frequency resolutions comparing to STFT spectrum; by analyzing the Hilbert marginal spectrum, the main frequency of the leakage signal was determined; this HHT method can be successfully applied in acoustic emission testing of valve leakage.
出处
《振动与冲击》
EI
CSCD
北大核心
2012年第23期62-66,共5页
Journal of Vibration and Shock
基金
国家自然科学基金资助项目(51175485)
山东省自然科学基金资助项目(ZR2009FM011)
