基于变分模态分解的延时识别法在短延时微差爆破中的应用

短延时微差起爆时间的精确识别，对复杂环境下隧道爆破优化设计和减振控制具有重要意义.针对经验模态分解（empirical mode decomposition，EMD）识别方法和希尔伯特-黄变换（Hilbert-Huang transform）识别方法存在的模态混叠和识别精度低的缺陷，提出了基于变分模态分解（variational mode decomposition，VMD）的延迟识别方法.分别采用VMD识别法、HHT识别法和EMD识别法对八达岭长城站爆破实测振动信号进行处理.结果表明:在短延时爆破识别中，HHT法的识别结果波峰杂乱，无法精确找到起爆时间；VMD法和EMD法的识别率随着雷管延时的增加而增加；延时低于20 ms时，EMD法的识别率约为VMD法的3/4，延时高于20 ms时，两种方法识别率都在80%以上，以上结果表明VMD法在短延时爆破的优化设计中具有一定的参考价值. 

Application of Variational Mode Decomposition Based Delay Time Identification in Short Millisecond Blasting

The accurate identification of delay time in short millisecond blasting is of great significance to optimal design and vibration control of tunnel blasting. Based on the defects of empirical mode decomposition (EMD) identification method and Hilbert-Huang transform (HHT) identification method such as mode mixing and low detection precision, a delay time identification method based on variational mode decomposition (VMD), namely VMD identification method, was proposed. VMD identification method, HHT identification method and EMD identification method were respectively used to process the measured blasting vibration signals of Badaling Great Wall Station. The results show that: in the identification of short delay blasting, the crest of HHT method was messy, and the initiation time could not be identified accurately; the recognition rate of VMD method and EMD method increased with the increase of detonator delay time; when the delay time was lower than 20 ms, the recognition rate of EMD method was about 3/4 that of VMD method, and when the delay was higher than 20 ms, the recognition rate of both methods was above 80%. The above results show that VMD method is of certain reference value to the optimization design of short millisecond blasting.