基于随机采样的频率域多路径波场模拟与偏移成像  

Multipath seismic simulation and imaging in frequency domain based on random sampling method

作　　者：李博[1,2] 

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 [2]中国科学院地质与地球物理研究所油气资源研究重点实验室,北京100029

出　　处：《石油物探》2017年第3期382-389,438,共9页Geophysical Prospecting For Petroleum

基　　金：国家重点研发计划项目(2016YFC0601101)资助~~

摘　　要：采用射线追踪方法的常规Kirchhoff深度偏移不能全面描述焦散区的波现象,有可能导致偏移成像产生畸变,利用此类方法进行复杂介质条件下的偏移成像,不可避免地存在焦散问题。详细推导了二次震源的波场模拟理论方法与区域分解的实施方案,在确保局部空间不存在射线交叉的前提下进行区域分解;基于惠更斯二次震源理论进行波场外推,逐步完成整个区域的波场延拓。提出了基于随机稀疏采样与低秩分解的波场模拟高效算法,借助于GPU的计算能力实现了频率域多路径波场模拟与成像,大幅提高了计算效率。数模实验结果表明,该方法能正确处理焦散问题,实现多路径情况下反射界面的正确成像。Conventional Kirchhoff prestack depth migration cannot handle the wave fields with caustic phenomenon; consequently, the subsurface reflector imaging could be distorted.In the complex media,Kirchhoff migration based on ray tracing must be inevitably affected by the caustic problem.In this paper,a multipath wave field simulation method based on the secondary source Huygens theory in complex media is proposed.Firstly, we ensure that there is no ray-cross in the local scope for domain decomposition. Then we use the secondary source Kirchhoff-Huygens integral method to gradually complete wave field extrapolation for the entire imaging area. In this way, we can correctly handle the caustic phenomena and get the correct imaging reflector.The wave field simulation method and domain decomposition strategy of secondary sources definition we described here have included large-scale matrix calculation that is the main bottleneck of efficiency. We propose a wavefield simulation algorithm based on random sampling and low rank decomposition method for large-scale matrix multiplication. On the other hand, we use GPU as the computing devices to realize the multipath seismic simulation and imaging in frequency domain, which greatly improve the computational efficiency. The numerical simulation test results show caustic could be removed and the correct reflector imaging is realized with this method.

关 键 词：偏移成像 波场外推 焦散 格林函数 低秩分解 随机采样 

分 类 号：P631[天文地球—地质矿产勘探]