复杂山地层析反演静校正新方法及应用  

A new tomographic inversion for static corrections in complex mountain areas

作　　者：曾庆才[1] 曾同生[1] 欧阳永林[1] 王兴[1] 宋雅莹 胡新海[1] 

机构地区：[1]中国石油勘探开发研究院廊坊分院,河北廊坊065007

出　　处：《石油地球物理勘探》2017年第3期418-425,441,共9页Oil Geophysical Prospecting

基　　金：国家油气重大专项"致密气富集规律与勘探开发关键技术"之课题"致密气有效储层预测技术"(2016ZX05047-002)资助

摘　　要：中国中西部前陆盆地油气资源丰富,但复杂的近地表结构及难以辨识的初至导致静校正问题异常突出,这已成为制约前陆盆地地下构造准确成像的一大瓶颈。为了解决这一问题,本文提出了一种复杂山地层析反演静校正方法:首先基于小波变换和Lipschitz指数获取有效初至信号并采用相关算法拾取准确的初至,然后基于BP神经网络和LSQR的双尺度反演算法求解精细的近地表速度模型,最后计算精确的静校正量。将该方法应用于塔里木库车克深地区实际地震资料,结果表明该方法有效提高了复杂山地的静校正精度。Foreland basins in central and western regions of China are rich in oil and gas resources, but their complex nearsurface structures which cause illegible first break lead to serious statics corrections problem, and hamper accurate imaging of subsurface geologic structures. To tackle this problem, we propose in this paper a new tomographic inversion for static corrections. First, we obtain valid first breaks based on wavelet transform and Lipschitz exponent, and extract automatically first break times with the correlation algorithm. Then we search for the global optimum solution of nearsurface structures with a dual-scale inversion approach by combining BP neural network and LSQR. Finally, we calculate precise statics. Our tests on real seismic data from Keshen area in Ta-rim Basin prove the effectiveness of the proposed method. © 2017, Editorial Department OIL GEOPHYSICAL PROSPECTING. All right reserved.

关 键 词：静校正 小波变换 Lipschitz指数 LSQR BP神经网络 

分 类 号：P631[天文地球—地质矿产勘探]