论文部分内容阅读
人工源极低频电磁法因其具有场源长度大,观测范围广的特点,使得极低频电磁波的传播会受到电离层和位移电流的影响。现阶段针对该方法的三维正反演研究尚处于起步阶段。作为一项探索性的尝试,本文首先给出了电离层、空气以及地下介质耦合情况下一维极低频背景电磁场的计算方案,并对电离层影响下背景电磁波的传播特征进行了分析、总结。通过将之与求解二次电场的交错网格有限差分数值模拟算法整合,实现了人工源极低频电磁法的三维正演。针对人工源极低频探测中可能遇到的近区、过渡区数据反演问题,本文进而采用针对该方法的三维有限内存Broyden-Fletcher-GoldfarbShanno(Limited-memory BFGS,LBFGS)带源反演算法,实现了对全区张量阻抗数据的直接反演。文中详细介绍了目标函数梯度计算这一LBFGS反演中的核心问题。合成数据反演算例结果表明在LBFGS反演中,选择恰当的近似Hessian矩阵能够有效提高反演效率。高低阻异常同时存在下的反演模型响应告诉我们张量阻抗反对角元素对恢复地下电性结构的贡献远大于主对角元素。与常规标量数据反演相比,张量数据反演在异常体的恢复和背景电阻率的控制方面具有明显的优势。
Due to its characteristics of large field source length and wide observation range, artificial low frequency electromagnetic method makes the transmission of very low frequency electromagnetic wave affected by ionosphere and displacement current. At this stage of the method for the three-dimensional forward and inverse analysis is still in its infancy. As an exploratory attempt, this paper first gives a calculation scheme of one-dimensional very low frequency background electromagnetic field under the coupling of ionosphere, air and subterranean media, and analyzes and summarizes the background electromagnetic wave propagation under the influence of ionosphere. By integrating this method with the finite difference numerical simulation algorithm of staggered grid for solving the secondary electric field, the three-dimensional forward modeling of artificial source and low-frequency electromagnetic method is realized. In order to solve the problem of near-area and transition-region data inversion in artificial low-frequency sounding, a three-dimensional finite-memory finite-memory BFGS (LBFGS) Realize the direct inversion of tensor impedance data in the whole area. In this paper, the core problem of LBFGS inversion in objective function gradient calculation is introduced in detail. The result of the synthetic data inversion shows that in the LBFGS inversion, choosing the approximate approximate Hessian matrix can effectively improve the inversion efficiency. The inverse model response to simultaneous presence of high and low resistivity anomalies tells us that the contribution of anti-diagonal elements of tensor impedance to the restoration of the subsurface electrical structure is much larger than that of the main diagonal elements. In contrast to conventional scalar data inversion, tensor inversion has obvious advantages in the recovery of anomalous bodies and the control of background resistivity.