Magnetotelluric 3D Forward Research Based On The Aggregation-based Algebraic Multigrid Algorithm

With the development of electronic technology, 3D electromagnetic method of data collection has been widely carried out. But because of slow calculation and instability of algorithm, the three-dimensional electromagnetic forward modeling and inversion technology practical application is not common.. Seriously hampered the development of MT methods. Traditional 1D and 2D MT inversion is difficult to clearly reflect the underground complex geological structure, so 3D inversion of MT usefulness imminent.Looking for an efficient and stable algorithm is one of the key factors driving the rapid development of the MT 3D inversion. This paper study 3D MT based on ggregation-based algebraic multigrid algorithm(AGMG) Numerical Simulation according to MT 3D forward the existing problems.In this paper, for 3D MT method forward questions, from the MT method satisfies Maxwell’s equations, staggered grid finite difference in time, combined with mixed boundary conditions for the formation of large sparse linear equations. And then using modern computational mathematics in the best one technique for solving linear equations aggregation-based algebraic multigrid algorithm(AGMG) to solved. Speed up the process of solving the three-dimensional electromagnetic method being played. We use graph theory thought, judging by the strength of the virtual node connectivity features, to achieve smooth aggregate matrix nonzero elements(coarsening) and Gauss-Seidel smooth iterations and different cycle iterations technology, to solve MT 3D forward equation.This paper solved the 3D Laplace equation, compare the advantages AGMG algorithm. It shows AGMG algorithm with respect to the traditional iterative algorithm convergence, good stability, iterative error linear attenuation characteristics. Analysis AGMG method factor influence the outcome of the algorithm, optimum solution parameters.In addition to using AGMG serial algorithm, but also introduced AGMG parallel algorithms and take advantage of computing resources to speed up the process forward. Then based MT method, for example, by a vertical fault model forward modeling,verification AGMG algorithm correctness, efficiency and stability of the three-dimensionalin numerical simulation of the MT. Finally analysing a typically electric model( low resistance model, high resistance model and combined model) different meshed proved AGMG adaptive algorithm for different models, can achieve fast convergence solution accuracy. Therefore, this article being developed by MT 3D algorithm with high accuracy,fast convergence, iterative stability, etc., to provide reference for the MT 3D forward modeling and inversion of practical promotion.