Two-dimensional Element Free Galerkin Method Forward Modeling For Magnetotelluric

Among the MT forward modeling methods, the finite element method is need to divide a large number of grid cells, the form of grid cell and size has a direct impact on the calculation accuracy, and the computation time is every large, and not suited to mesh distortion occurs due to large deformation. The paper studies a new MT forward modeling algorithm-Element Free Galerkin Method (EFG). EFG only needs simply node, no units, and gets rid of the bondage of unit. Applying the node information and the weight function in its local support domains can realize local precise approach, the shape function and its derivative has the commendable consistency, while the shape function is structured by the Moving Least Squares (Moving Least Squares, MLS); that can make the EFG with high precision, strong adaptive ability, and large flexible features.This paper reviews the MT forward modeling, with its development history, the research present situation, the theory of EFG and its development history. Then the paper detailed deduces the MT boundary problem and variational problem from Maxwell equations, Then it gives the detailed introduction on the basic principle of EFG, and that mainly discusses on the EFG with the structure of the shape function, the selection of the weight function, the influence domain, the gauss integral grid division and its realization methods and measures. And then analyzes the advantages of using EFG in the MT forward modeling. And detailed deduces how to deal with essential boundary conditions with the penalty function method and the Lagrange multipier method. I write the EFG program use the Moving Least Squares Method to structured the shape function, cubic spline function as the weight function, choose four gauss integeral formula for numerical integration and using penalty function to deal with boundary conditions. Use this program I make forward modeling on a series of typically electric model of MT. For the one-dimensional horizontal layered model, comparing the EFG forward results that with the analytical solution, it proved the EFG program is effective. Make EFG forward for some simple2-D model of MT, obtained some related conclusions.42pictures,3sheets,60references.