大地电磁非线性共轭梯度拟三维反演

提出了非线性共轭梯度法大地电磁拟三维反演.该方法选取共轭梯度反演算法为拟三维反演的核心.在计算灵敏度（Jacobian）矩阵时，吸取近似灵敏度矩阵思想，采用一维灵敏度矩阵来代替三维灵敏度矩阵，并对非测点的灵敏度元素提出一种近似方法.在第一次反演之后，采用拟牛顿法更新灵敏度矩阵.拟三维反演法在很大程度上节省了计算时间，并且理论模型和实际资料的反演试算结果表明大地电磁拟三维反演法具有一定的实用价值.

Pseudo-Three-Dimensional magnetotelluric inversion using nonlinear conjugate gradients

We propose nonlinear conjugate gradient pseudo-3-D method was selected as the inversion kernel. When calculating MT inversion algorithm. The conjugate gradient sensitivities matrix, instead of computing 3-D sensitivities, we attempt to compute 1-D sensitivities. Since the sensitivity elements of non-stations are zero, which should be non-negligible values, we adopt a method of approximation. We update sensitivities matrix by the quasi-Newton method after the first inversion. Psudeo-3-D MT inversion can save computational time greatly. Inversion results of a synthetic data set and real data set show that the psudeo-3-D MT inversion has some practical applications.