摘要
针对现有矿山微震监测系统信号自动识别难的问题,提出基于经验模态分解(EMD)和形态分形维数的识别方法。首先,采用EMD将原始信号分解为若干个本征模态分量(IMF),选择前5个分量进行重构得到新的信号。其次,求出处理后信号的形态学分形维数,利用微震波形和爆破波形分形维数的差异进行信号识别。对50组微震波形和50组爆破波形进行试验研究。对比未经EMD处理的形态学分形维数以及经EMD处理的盒维数识别结果。研究结果表明:50组微震波形和50组爆破波形在形态学分形维数为1.4时具有较高的识别率;微震波形维数主要在1.4以下,爆破波形维数则基本高于1.4;EMD结合形态学分形维数的识别效果最好,为微震监测波形识别提供了新途径。
Considering that it is difficult to recognize signals automatically in microseismic monitoring system of mine, the new recognition method was proposed based on empirical mode decomposition （EMD） and morphological fractal dimension. Firstly, origined-signal was decomposed to some intrinsic mode ftmctions （IMF） by EMD, and new signal was obtained by adding former five components, Then, morphological fractal dimension of the new signal was calculated to recognize microseismic waveform and blasting waveform. The 50 microseimic waveforms and 50 blasting waveforms were tested. Finally, the dimensions were calculated with two different methods, i.e. morphological fractal dimension without EMD and fractal box dimension with EMD. The results show that 50 microseimic waveforms and 50 blasting waveforms show high recognition ratio in morphological fractal dimension of 1.4. Dimensions of microseismic waveform and blasting waveform are mainly below and over 1.4, respectively. The new recognition method has the best recognition result, and provides a new way for waveform recognition.
出处
《中南大学学报（自然科学版）》
EI
CAS
CSCD
北大核心
2017年第1期162-167,共6页
Journal of Central South University:Science and Technology
基金
国家自然科学基金资助项目(51374244)~~
关键词
微震
EMD
形态学
分形维数
波形识别
microseimic
EMD
morphology
fractal dimension
waveform recognition
作者简介
通信作者：赵国彦，教授，博士生导师，从事采矿工程、矿山安全和岩石力学与工程等研究；E-mail：810703752@qq．com
