应用于大规模三维数据反演的拟线性近似方法的计算精度和应用范围至今仍是一个比较模糊的概念.本文首先实现了基于拟线性近似方法(对角拟线性近似、标量拟线性近似、拟解析近似和局部拟线性近似)的三维电磁场数值模拟,然后通过正演计算,对这些方法的计算精度进行了系统的对比研究.理论研究结果表明:对角拟线性近似方法精度最高,适用范围最广,对电性变化在3~4个数量级内的地电模型都能给出精确的计算结果;拟解析近似方法计算精度只低于对角拟线性近似方法,它适用于异常电导率与背景电导率比值为2~3个数量级的地电模型;标量拟线性近似方法和局部拟线性近似方法的精度较低,只适用于异常电导率与背景电导率比值小于2个数量级的情况.%The accuracy and the application scope of quasi-linear approximation methods, including scalar quasi-linear approximation, diagonal quasi-linear approximation, quasi-analytic approximation and localized quasi-linear approximation, are still a relatively misty concept in large-scale three-dimensional data inversion.In the paper, we implemented three-dimensional numerical simulation based on these approximation methods.By forward modeling,we systematically compared the accuracy of simulation results, and constrained the range of applications of these methods.Theoretical results show that: The diagonal quasi-linear approximation method has the highest accuracy and the widest range of applications among these methods and can give accurate results when electrical resistivity changes within 3-4 orders of magnitude;The accuracy of the quasi-analytical approximation method is a bit lower than that of the diagonal quasi-linear approximation method, and is suitable for calculating the geoelectrical model whose conductivity contrast between abnormal and background is roughly a few times to a hundred times;The accuracy of scalar quasi-linear approximation method and localized quasi-linear approximation method is the lowest, and they are only able to be used to calculate the geoelectric model of anomalous conductivity and background conductivity ratio of roughly tens of times.
展开▼
