论文详情
微地震资料的处理、解释首先需要拾取精确的初至信息。微地震初至的人工拾取虽然精度较高,但工作量大,不能满足实时处理的需要。为此,提出了一种基于小波多尺度分解和高阶统计量相结合的长短时窗峰度比(wavelet transform based short time window kurtosis/long time window kurtosis,W-STK/LTK)微地震初至拾取方法。考虑到有效微地震信号频率较低,而噪声信号频率则相对较高,首先对微地震信号进行小波多尺度分解,剥离有效信号与噪声;接着对分解得到的最大尺度信号应用基于高阶统计量的长短时窗峰度比(STK/LTK)算法;最后在分析特征曲线异常点特征的基础上识别微地震有效信号并拾取初至。模型数据和实际资料测试结果表明,该方法能够从信噪比较低的微地震资料中较准确地拾取微地震P波初至。
Accurate first arrival information is necessary for the processing and interpretation of microseismic data.Although manual picking has a high accuracy,it is difficult to satisfy the real-time processing because of its poor efficiency.Therefore,we proposed a wavelet multi-scale decomposition and high-order statistics based short time window kurtosis / long time window kurtosis (W-STK/LTK) microseismic first arrival picking method.By considering the low frequency of effective microseismic signals and relatively high frequency of noise,we firstly carry out multi-scale wavelet decomposition on microseismic signals to strip significant signals and noise.Then,we apply high-order statistics based STK/LTK algorithm on the maximum-scale signals achieved by decomposition.Finally,the microseismic significant signals are identified and the first arrival is picked up based on the analysis of the abnormal points on characteristics curve.Model and actual data testing results indicate that the method can be adopted to accurately pick up the P-wave first arrival from the microseismic data with relatively low S/N.
国家高技术研究发展计划(863计划)项目(2011AA060303)和中央高校基本科研业务费专项资金(13CX02098A)联合资助。