岩石破裂微震与爆破振动信号时频特征提取及识别方法

发布时间：2018-08-09 14:23

【摘要】：微震信号蕴藏着丰富的岩体破裂信息,对其监测并进行数据处理分析可以获取岩体破裂的位置及能量释放情况,目前已在冲击地压、煤与瓦斯突出等煤岩动力灾害监测预警领域得到广泛应用。但是矿下环境复杂多变,需要经常进行岩石爆破作业,拾振器拾取的微震信号中往往掺杂着无法识别的爆破干扰信号,影响微震监测及定位结果。因此如何有效的提取两者的特征参数信息来识别岩石破裂微震信号和爆破振动信号显得尤为重要。本文基于岩石破裂微震信号和爆破振动信号的时变非平稳特征,通过对比几种时频分析的方法性能—短时傅里叶变换、小波变换和希尔伯特黄变换,提出了基于集合经验模态分解的岩石破裂微震信号和爆破振动信号的时频能量特征提取和识别方法。首先,通过小波阈值去噪,将待测信号中的噪声干扰成分尽可能的剔除,真实地还原信号波;其次对去噪的待测信号进行集合经验模态分解(EEMD),获得一系列本征模态函数(IMF);最后求得每个IMF的能量占总信号能量的比例来作为待测信号的时频能量分布。由于岩石破裂微震信号和爆破振动信号的频率分布状况不同,故将求得的本征模态函数能量比值的分布情况来作为其特征参数,来识别岩石破裂微震信号和爆破振动信号。通过对80组典型的煤岩破裂微震信号和爆破振动信号进行实验,结果显示,煤岩破裂微震和爆破振动信号IMF能量分布有较大差别,煤岩破裂微震信号主要集中在IMF2、IMF3和IMF4的20-100Hz低频段,爆破振动信号则在IMF1的225-375Hz高频处较为集中。为把两者信号差异最大化,从而形成区分两者的有效特征参数,将IMF2、IMF3和IMF4频段能量合并为新频段,爆破振动信号在IMF1与煤岩破裂微震信号在IMF(2+3+4)频段内能量值所占比例均在80%以上,区别最为明显,故将IMF1与IMF(2+3+4)能量特征比例作为区分煤岩破裂微震信号和爆破振动信号的特征指标。该分析方法为煤矿识别微震信号事件和爆破信号事件提供了一种新的思路,利用两者能量分布差异较大、特征对比明显等特点,可以实现对两类波形信号的有效辨识。
[Abstract]:The microseismic signal contains abundant information of rock mass rupture. Monitoring and data processing can obtain the position of rock mass rupture and energy release. At present, rock burst is already under ground pressure. Coal and gas outburst are widely used in the field of monitoring and warning of coal and rock dynamic disasters. However, the mining environment is complex and changeable, so rock blasting is often needed. The microseismic signals picked up by vibration pickers are often mixed with unidentifiable blasting interference signals, which affect the monitoring and positioning results of microearthquakes. Therefore, it is very important to extract the characteristic parameter information to identify the rock rupture microseismic signal and the blasting vibration signal. Based on the time-varying non-stationary characteristics of rock rupture microseismic signals and blasting vibration signals, this paper compares the performance of several time-frequency analysis methods-short time Fourier transform, wavelet transform and Hilbert-Huang transform. A method of extracting and identifying time-frequency energy features of rock rupture microseismic signal and blasting vibration signal based on set empirical mode decomposition is proposed. Firstly, the wavelet threshold is used to remove the noise interference from the signal to be tested, and the signal wave can be truly restored. Secondly, the set empirical mode decomposition (EEMD),) of the denoised signal is used to obtain a series of intrinsic mode functions (IMF);). Finally, the ratio of the energy of each IMF to the total signal energy is obtained as the time-frequency energy distribution of the signal to be tested. Because the frequency distribution of rock rupture microseismic signal and blasting vibration signal is different, the distribution of eigenmode function energy ratio is taken as its characteristic parameter to identify rock rupture microseismic signal and blasting vibration signal. Based on the experiments of 80 typical coal and rock rupture microseismic signals and blasting vibration signals, the results show that the IMF energy distribution of coal and rock rupture microearthquakes and blasting vibration signals is quite different. The microseismic signals of coal and rock fracture are mainly concentrated in the low frequency band of 20-100Hz of IMF2F3 and IMF4, while the vibration signals of blasting are concentrated at the high frequency of 225-375Hz of IMF1. In order to maximize the difference between the two signals and form an effective characteristic parameter to distinguish the two, the energy of IMF2F3 and IMF4 band is combined into a new frequency band. The proportion of blasting vibration signals in the IMF (234) frequency band of IMF1 and coal rock rupture microseismic signals is more than 80%, the difference is most obvious. Therefore, the ratio of energy characteristic of IMF1 and IMF (234) is taken as the characteristic index to distinguish the microseismic signal of coal and rock fracture from the vibration signal of blasting. This analysis method provides a new way of thinking for coal mine to identify microseismic signal events and blasting signal events. The two kinds of waveform signals can be effectively identified by using the characteristics of great difference of energy distribution and obvious characteristic contrast between the two kinds of signals.
【学位授予单位】：山东科技大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TD326

【参考文献】