冻结立井爆破振动信号去噪研究  

Research on Blasting Vibration Signal Denoising Method in Freezing Vertical Shaft

作　　者：付晓强[1] 陈岐范 李明[1,2] 左进京[1] 史国利 

机构地区：[1]中国矿业大学(北京)力学与建筑工程学院,北京100083 [2]兖矿集团有限公司,山东邹城273500

出　　处：《煤矿安全》2017年第9期47-50,共4页Safety in Coal Mines

基　　金：国家重点研发计划资助项目(2016YFC0600903)

摘　　要：针对冻结立井监测到的爆破振动信号的高噪声、短持时和非线性的特征,提出了经验模态分解-去趋势项波动分析(Empirical Mode Decomposition-Detrended Fluctuation Analysis,EMDDFA)组合去噪方法。采用4项指标综合评价了EMD-DFA法、EMD和EEMD法、小波阈值法和小波熵去噪法去噪效果。结果表明:被强噪声污染的爆破振动信号采用EMD-DFA组合法去噪得到了信噪比和互相关度最高,相应的误差最小,去噪后的信号保留了爆破信号的瞬态非平稳特征,去噪效果最优。According to the characteristics of high noise, short duration and nonlinear of freezing vertical shaft blasting vibration signal in practical engineering, the method of EMD-DFA （Empirical Mode Decomposition-Detrended Fluctuation Analysis） combining de-noising method is proposed. Four indexes were used to evaluate the de-noising effect of EMD-DFA, EMD and EEMD, wavelet threshold and wavelet entropy. The results show that： the blasting vibration signals including strong noise pollution by using the combined method of EMD-DFA de-noising, SNR and cross-correlation is the highest, the corresponding error is minimum, the de-noising signal retains the transient characteristics of non-stationary signal blasting, de-noising effect is optimal.

关 键 词：经验模态分解 去趋势波动分析 爆破振动 信号去噪 时频分析 

分 类 号：TD235.1[矿业工程—矿井建设]