摘要
为解决轧机扭振非平稳瞬态冲击信号的瞬态时频分析难题,提出了基于EMD和Hilbert-Huang变换的瞬态振动信号时频分析方法.利用EMD分解提取信号的固有模态函数(IMF),再结合Hilbert变换,求解瞬时频率,进而得到信号的Hilbert-Huang时频谱及边际谱,从而提取扭振信号瞬态特征.通过仿真实验验证了该方法的可靠性,通过轧机在咬钢抛钢时实测瞬态信号的分析,表明了该方法的可行性.
To solve the problem of extracting the transient characteristics of torsional vibration nonstationary transient impact signal of the rolling mill, a novel method was presented using signal processing theory. First, the intrinsic mode function（IMF） of the signal was extracted via the empirical mode decomposition（EMD）. Then combined with Hilbert transformation, the extracted model of the instantaneous frequency of single-frequency vibration transient signal was established to solve the instantaneous frequency. Afterwards, the Hilbert spectrum and the Hilbert boundary spectrum were obtained. By this means, the transient characteristics of the torsional vibration signal can be effectively extracted. Compared with wavelet decomposition, this algorithm avoids frequency overlapping in the wavelet decomposition and has the advantage of time scale self-adaption. It overcomes the difficulty of selecting elementary wavelet. The numerical simulation shows the feasibility of the method. When roiling mills grip or throw the steel, the example analysis of transient signal show the reliability of the method.
出处
《苏州市职业大学学报》
2009年第1期37-40,共4页
Journal of Suzhou Vocational University
关键词
瞬态冲击
特征提取
瞬时频率
EMD分解
Hilbert—Huang时频谱
transient impact, feature extraction, instantaneous frequency, empirical mode decomposition, hilbert-Huang time-frequency spectrum
