论文部分内容阅读
经验模式分解(EMD)具有高度的自适应性,但是存在较大的端点效应。为此提出利用支持向量机-镜像延拓法对其端点效应进行处理,实际效果良好。提出了爆炸微差延时的改进EMD时间尺度识别法,并以某工程中微差爆炸监测到的爆炸振动信号为例,通过改进的EMD时间尺度法对监测得到的爆炸振动信号进行分解,从而获得多个固有模态函数(IMF)分类;然后通过Hilbert变换提取这些IMF中占主成分的IMF分量的包络幅值,进而达到识别实际微差延时的目的。该方法为爆炸延期参数设计提供了分析手段。
Empirical Mode Decomposition (EMD) is highly adaptive, but has a large endpoint effect. To this end, we propose to use support vector machine - mirror continuation method to deal with its end effect, the actual effect is good. An improved EMD time-scale identification method was proposed for the differential delay of explosion. The explosion vibration signal monitored by the differential-error explosion in a project was taken as an example to decompose the monitored explosive vibration signal by the improved EMD time-scale method Then, the IMFs of these IMFs are extracted by Hilbert transform, then the envelope amplitude of the IMF components of these IMFs is extracted, so as to achieve the purpose of identifying the actual differential delay. The method provides an analytical tool for the detonation delay parameter design.