摘 要: 微震監測系統主要用來監測和分析各種生產及開采活動中由巖石破壞所引起的微震。微震監測原理與地震監測相同,主要包括數據采集,事件的信號識別和定位,其中信號識別主要是指P波震相的識別,它直接影響微震定位的可靠性。目前在地震監測中,廣泛使用的識別算法主要是pick_ew和FilterPicker算法。使用實際數據對這兩個算法進行研究并優化其參數,使它們能夠更好地應用在微震事件的監測中。由于pick_ew配置參數的復雜性,主要是依據前人使用的歷史經驗對參數進行優化;而FilterPicker算法只需要調整一個參數就可以將其更好的應用在微震監測中,主要使用0.618法來找出這個參數值。
關鍵詞: 微震監測; 自動識別; 參數優化; pick_ew; FilterPicker
中圖分類號: TN911.7?34 文獻標識碼: A 文章編號: 1004?373X(2013)23?0065?05
Research and parameter optimization on microseismic event signal recognition algorithm
WU Yi?qun, WU Guan?mao
(School of Computer Science and Technology, Anhui University of Science and Technology, Huainan232001, China)
Abstract: Microseismic monitoring system is mainly used to monitor and analyze the microseisms caused by rock destruction of various productions and mining activities. Microseismic monitoring principle is the same as earthquake monitoring, which includes data acquisition, signal identification and location of events. The signal identification mainly refers to the P?wave seismic phase identification, which directly affects the reliability of microseismic monitoring system. Currently the recognition algorithms widely used in seismic monitoring are pick_ew algorithm and FilterPicker algorithms. Actual data is used to study these two algorithms and optimize their parameters, to enable them to better use in monitoring microseismic events. Since the complexity of the parameter configuration of pick_ew, the historical experience of our predecessors are mainly used to optimize the parameters, while it only need to adjust one parameter in FilterPicker algorithm, that can make it better application in microseismic monitoring. The parameter value is found out by 0.618 method.
Keywords: microseismic monitoring; automatic identification; parameter optimization; pick_ew; FilterPicker
0 引 言
微震,局部地區的小型地震,可作為巖石破壞的標志,在工程施工和礦山安全領域得到越來越多的關注。近年來,微震監測系統得到了廣泛地應用。微震監測系統由數據采集,事件的信號識別,定位及相應的參數計算等模塊組成。在這些模塊中,事件的信號識別是最基本的模塊,它直接影響到定位及所有其他微震相關參數計算的準確性。與地震相比,微震發生的頻率要大得多,對每一微震事件進行人工識別是不切實際的,鑒于此,微震事件的自動識別技術一直得到人們的重視。
Allen提出了pick_ew算法[1]以及Baer和Kradolfer提出了BK87算法[2]用于地震事件信號的自動識別。這兩種信號識別算法都是通過比較信號特征函數的短期平均值(STA)、長期平均值(LTA)和閾值之間的關系。……