节点文献

双程声波方程叠前逆时深度偏移及其并行算法

Reverse-time depth migration of two way acoustic equations and its parallel algorithm

  • 推荐 CAJ下载
  • PDF下载
  • 不支持迅雷等下载工具,请取消加速工具后下载。

【作者】 何兵寿张会星韩月

【Author】 HE Bing-shou,ZHANG Hui-xing,Han Yue(Key Lab of Submarine Geosciences and Prospecting Techniques,Ministry of Education,Ocean University of China,Qingdao 266100,China)

【机构】 中国海洋大学海底科学与探测技术教育部重点实验室

【摘要】 从双程声波方程出发,在交错网格空间中推导了地震波逆时延拓的高阶有限差分算子,依据最佳匹配层的方程分裂思路得到了一阶声波方程的最佳匹配层(PML)吸收边界条件及其高阶差分格式。利用炮点波场对上、下行波互相关后的逆时偏移结果进行振幅校正,补偿了深层能量。在上述成果的基础上实现了声波方程的叠前逆时深度偏移。偏移算法的脉冲响应表明:逆时偏移算子具有良好的陡倾角地层成像能力与深层能量补偿能力。研究了基于消息传递接口(MPI)的声波方程逆时偏移并行算法,给出了适于并行求解的计算空间划分方法,实现了双程声波方程逆时偏移的粗粒度并行运算,有效地提高了双程声波方程逆时偏移的计算效率。算法对Marmousi模型理论数据的偏移成像得到了满意效果。

【Abstract】 Based on two-way acoustic wave equations,a high-order finite difference reverse-time propagation scheme was established using a high precision operator in staggered-grid.According to the splitting method of seismic wave equations,a perfectly matched layer(PML) absorbing boundary condition and its finite difference scheme were also proposed.The amplitude of migration result was corrected by the square of source wavefield to compensate the energy of deep stratum.Based on above techniques,the pre-stack reverse-time depth migration of compressional seismic wave was achieved in two dimensional space.The impulse response implies that the algorithm has the advantage of energy compensation and applicable for dip angle reflector imaging.In order to improve the processing speed,studied a parallel algorithm of reverse-time migration based on Message Passing Interface(MPI).The computation domain division method and communication scheme that were suitable for parallel algorithm was proposed.Using above algorithms,a coarse-grained parallel reverse-time migration algorithm was realized,which could improve the computation efficiency greatly,and the synthetic data processing of marmousi model obtained a satisfactory result.

【基金】 国家自然科学基金资助项目(40804021);国家自然科学基金重点资助项目(40839901)
  • 【文献出处】 煤炭学报 ,Journal of China Coal Society , 编辑部邮箱 ,2010年03期
  • 【分类号】P631.4
  • 【被引频次】18
  • 【下载频次】400
节点文献中: 

本文链接的文献网络图示:

本文的引文网络