Three‐Dimensional Rapid Relaxation Inversion for the Magnetotelluric Method
Three‐Dimensional Rapid Relaxation Inversion for the Magnetotelluric Method
复制标题
DOI：
10.1002/cjg2.442
复制
发表时间：
2003-11
期刊：
影响因子：
--
通讯作者：
H. Tan;Qinfan Yu;Booker John;Wenbo Wei
中科院分区：
文献类型：
--
作者：
H. Tan;Qinfan Yu;Booker John;Wenbo Wei
大地电磁法三维反演的关键是找到一种有效的方法来计算灵敏度矩阵。通过对三维阻抗表达式的进一步分析，导出了三维Frechet核函数的计算公式，这是最常用的核函数，其形式与二维情况下的核函数相似。通过定义目标函数对三维电性结构进行光顺处理，得到了三维快速弛豫反演（RRI）的最小结构迭代过程。三维RRI反演具有二维RRI反演的所有优点。在加有高斯噪声的合成数据集上对3-D RRI算法进行了测试。实验模型为二维棱镜和三维棱镜。反演的电阻率模型与真实模型拟合较好。本文还利用日本茅部地区地热资源勘探数据集对三维RRI算法进行了测试。实验结果表明，本文提出的三维RRI反演算法速度快，稳定性好，能在PC机上有效地运行。
The crucial step of doing 3-D inversion for the magnetotelluric method is to find an effective way to calculate the sensitivity matrix. Through analyzing the expressions of 3-D impedances further, the formulas of 3-D Frechet kernel functions are derived, which are the most common ones and are similar to the ones in forms used in 2-D case. Defining the object function to smooth the 3-D electrical structures, we finally get the iteration procedure of 3-D Rapid Relaxation Inversion (RRI) with the minimum structure. The 3-D RRI inversion has all excellent points shown in 2-D RRI inversion. The 3-D RRI algorithm is tested on the synthetic data set with added Gaussian noise. The tested models are 2-D prism and 3-D prism. The inverted resistivity models fit the true models well. The 3-D RRI algorithm is also tested on Kayabe data set, which is collected in the Kayabe area of Japan at the purpose of prospecting geothermal resource. All tested results show that we've successfully developed a practical 3-D RRI inversion algorithm, which is fast, stable and can run effectively on PC.