基于共轭梯度法和互相关的最小二乘逆时偏移及应用（英文）

发布时间：2018-09-10 16:30

【摘要】：常规逆时偏移可以实现较好的构造成像,但由于照明不均等因素使得该方法不能实现对岩性储层的精确刻画。为了得到可靠的地下反射界面的反射系数,需要用反演的方式解决成像的问题。最小二乘逆时偏移(LSRTM)被称为线性反射率反演,它通过引入Hessian矩阵实现相对的高分辨率振幅保真成像。共轭梯度算法是非常高效的迭代算法,使得LSRTM方法变得实用。基于模型数据与观测数据的互相关程度判定速度模型的准确度及计算模型更新量,可以使得LSRTM摆脱地震子波的依赖,增强稳健性。从模型试算及实际资料处理中可以看出,相比常规RTM和单程波偏移方法,LSRTM的成像结果可以直接应用到后续的储层描述和四维地震中。本论文主要研究了最小二乘RTM的一阶近似,也就是线性Born近似。当遇到更复杂的地质构造时,可以通过考虑更高阶的近似来提高其应用效果。
[Abstract]:Conventional inverse time migration can achieve better structural imaging, but because of the uneven illumination, the method can not accurately depict lithologic reservoirs. In order to obtain the reflection coefficient of the underground reflection interface, the problem of imaging needs to be solved by inversion. Least square inverse time migration (LSRTM) is called linear reflectivity inversion. It uses Hessian matrix to realize relative high resolution amplitude fidelity imaging. Conjugate gradient algorithm is a very efficient iterative algorithm, which makes the LSRTM method practical. Based on the degree of cross-correlation between model data and observation data, the accuracy of velocity model and the calculation of model update can make LSRTM get rid of the dependence of seismic wavelet and enhance the robustness. Compared with conventional RTM and one-way wave migration methods, the imaging results of LSRTM can be directly applied to the subsequent reservoir description and four-dimensional earthquakes. In this paper, the first order approximation of the least square RTM, that is, the linear Born approximation, is studied. When a more complex geological structure is encountered, the application effect can be improved by considering higher order approximation.
【作者单位】：中国石油大学(华东);海洋国家实验室海洋矿产资源评价与探测技术功能实验室;
【基金】：sponsored by The National Natural Science Fund(No.41574098) Sinopec Geophysical Key Laboratory Open Fund(No.wtyjy-wx2016-04-2)
【分类号】：P631.4

【相似文献】