基于非结构化网格的大规模磁测数据正则化反演研究  

Study on the Regularization Inversion of Large Scale Magnetic Survey Data Based on Unstructured Grid

作　　者：刘阳 张志勇 LIU Yang;ZHANG Zhiyong(School of Geophysics and Measurement and Control Technology,East China University of Technology,Nanchang,Jingxi 330013,China)

机构地区：[1]东华理工大学地球物理与测控技术学院,江西南昌330013

出　　处：《铀矿地质》2023年第5期831-842,共12页Uranium Geology

基　　金：国家自然科学基金项目(编号:42164008);江西省自然科学基金项目(编号:2019BAB202012)联合资助。

摘　　要：磁测数据解析正演的内存需求与模型规模、观测点数成正比,以其为依托开展大规模磁测数据反演时,计算成本较大。文章利用非结构化网格有限单元法进行磁测正演,分析了地表观测网格、边界节点数和局部网格密度等因素对正演计算精度的影响;在此基础上,基于灵敏度矩阵隐式存储的高斯-牛顿法求解正则化反演目标函数,实现了大规模磁测数据三维反演。理论模型试算和铀矿区域实测数据案例表明,文章所述方法可保证正演计算的精确度,提高了计算效率、减少了反演的内存消耗,可用于复杂条件下大规模磁测数据的三维反演。The memory requirement of analytical forward modeling of magnetic survey data is proportional to the size of the model and the number of observation points,and the calculation cost is relatively high when large-scale magnetic survey data inversion is carried out based on it.In this paper,unstructured grid finite element method was used to carry out magnetic forward modeling,and the influence of surface observation grid,boundary node number and local mesh refinement on the accuracy of forward modeling was analyzed.On this basis,the Gauss-Newton method based on implicit storage of sensitivity matrix was used to solve the objective function of regularization inversion,and realize the three-dimensional inversion of large-scale magnetic measurement data.Theoretical model trials and measured data cases showed that the proposed method could ensure the accuracy of forward calculation,improve the calculation efficiency and reduce the memory consumption of inversion,therefore,it can be used for three-dimensional inversion of large-scale magnetic measurement data under complex conditions.

关 键 词：磁法勘探 磁化率 有限单元法 正则化反演 

分 类 号：P631.2[天文地球—地质矿产勘探]