高斯束偏移与高斯束层析反演速度建模  

Gaussian beam operator-based migration and tomography

作　　者：蔡杰雄[1] CAI Jiexiong(Sinopec Geophysical Research Institute,Nanjing 211103,China)

机构地区：[1]中国石油化工股份有限公司石油物探技术研究院,江苏南京211103

出　　处：《石油物探》2018年第2期262-273,共12页Geophysical Prospecting For Petroleum

基　　金：国家科技重大专项(2016ZX05014-001-002)资助~~

摘　　要：当前的层析反演速度建模方法大多基于射线传播算子,以射线长度为层析反演敏感度核函数,对复杂构造偏移成像的适应性不强,反演精度有待提高。基于波动方程的一阶Born近似和Rytov近似,从高斯束偏移成像条件出发,推导了成像域波动方程线性化走时层析反演核函数,该核函数的本质是有限频核函数,可通过高斯束积分表达的格林函数计算得到。利用该核函数替换常规射线层析核函数能提高层析反演精度。发展的高斯束层析反演速度建模方法通过方位-反射角度域共成像点道集与高斯束偏移串联并迭代实现复杂构造成像,该技术路线实用化程度高,能进一步提高当前工业界广泛应用的常规射线层析及射线偏移的成像精度,尤其是改善了低信噪比资料的成像质量。数值计算及实际数据应用证明了基于高斯束算子的偏移成像与成像域走时层析方法的有效性。Conventional ray-based travel time tomography for velocity modeling,regarding ray length as the sensitivity kernel,has limitations,low spatial resolution and low adaptability in complex heterogeneous media because of caustics.Starting from the Gaussian beam migration(GBM)imaging condition,a new travel time tomography sensitivity kernel is presented based upon the first order Born and Rytov approximation.This kernel is a finite-frequency kernel function;it can be obtained by the Green function,integrated by Gaussian beams.Using this new kernel could improve the accuracy of tomography inversions.The joint iteration process of Gaussian beam operator-based migration,a common-image gather extracted in the azimuth-reflection angle domain,and tomography for velocity modeling may improve image quality,especially for low SNR seismic data.Synthetic and field data tests verified the effectiveness and applicability of the proposed technique.

关 键 词：高斯束 成像域 走时层析 敏感度核函数 偏移 角度道集 

分 类 号：P631[天文地球—地质矿产勘探]