变分模态分解在爆破信号趋势项去除中的应用

爆破工程中，信号趋势项的准确去除对提高爆破振动信号分析的精度具有重要意义。针对经验模态分解（empirical mode decomposition，EMD）识别法存在的模态混叠和端头效应等缺陷，提出了基于变分模态分解（variational mode decomposition，VMD）去除信号趋势项的方法，即VMD法。叙述了VMD法识别爆破信号趋势项原理，并进行了仿真实验，结果表明:趋势项频率对分解效果的影响相对较小，当趋势项频率处于1～5 Hz之间时，频率对分解效果的影响基本保持不变；振幅对分解效果影响显著，且振幅越小，VMD法的分解效果越差。当趋势项振幅超过原始爆破信号最大振幅的1/3时，VMD法分解效果较好。最后，应用VMD法和EMD法对含有趋势项的实测爆破振动信号进行处理，认为相比于EMD法，VMD法处理后的信号基本一致且不存在端点效应，在爆破信号趋势项去除领域中具有更加广泛的适用性。

Application of variable mode decomposition in the removal of blasting signal trend items

Accurate removal of signal trend items is of practical importance for improving the accuracy of blasting vibration signal analysis. Here, aiming at the defects of EMD identification, such as mode aliasing and terminal effect, a method based on variational mode decomposition (VMD) to remove signal trend term is proposed. The principle of identifying the trend term of blasting signals by VMD method is described in details, and the simulation experiment was carried out. The results show that the influence of the trend term frequency on the decomposition effect is relatively small. When the trend term frequency is between 1 and 5 Hz, the effect of the frequency on the decomposition effect remains basically the same. The amplitude has a significant influence on the decomposition effect. Furthermore, the amplitude is smaller, the decomposition effect of the VMD method is worse. When the amplitude of the trend term exceeds 1/3 of the maximum amplitude of the original blasting signal, the VMD method has a better decomposition effect. Finally, the VMD method and the EMD method are applied to process the measured blasting vibration signal containing the trend term. Compared with the EMD method, the signals processed by the VMD method are basically consistent and have no terminal effect, and have wider applicability in the field of blasting signal trend item removal.