Focusing Inversion Images Of Magnetotelluric Data

The magetotelluric sounding is an excellent geophysical tool by measuring thesurface electromagnetic response of earth and analyzing the character ofelectromagnetic field to study the rock electricity and character of conformation. Theprinciple of MT is deducing the geoelectrical structures by analysing theelectromagnetic frequency response of earth. MT is used to study the structuralfeature of earth crust and upper mantle for the excellence that needn't the artificialelectromagnetic field and can get the deep information of earth. However, with thewide practical application and in-depth theory research, lots of new problems arepresented and need to be solved. Sponsored by CAS and Sinopec, focusing inversionimages method is presented and advantage of the method is proved. The maincontents and achievements of the thesis are following:(1) An efficient method named secondary finite element method (SFEM) ispresented to two-dimensional magnetotelluric numerical modeling, which solves forsecondary variations in the field. During deducing the rationle of the new procedurewe divide total field into primary and secondary field. Then we just need to solve thesecondary field equations for the primary field can get easily. The equations of thesecondary field are solved effortless by finite element method when the boundaryconditions are simple. The effect of resolution is analysed when we deal with theresponse of synthetic model by the secondary field and compare with the results usingtotal field FEM.(2) Usually a maximum smoothness stabilizing functional is used to stabilize theinversion process. The obtained solution is a smooth image by the traditionalstabilizer, which in many practical situations does not describe the examined objectproperly. During narrating the procedure of the approach, we analyse the character ofseveral traditional model object functions firstly, then introduce a new stabilizingfunctional named minimum support stabilizer and the new parametric functional ispresented. The simulated results show the program is right and this approach helps togenerate much more "focused" and resolved images of blocky geoelectrical structures than conventional MT inversion.(3) To speed up the calculation of the sensitivity matrix for 2-D magnetotelluricinversion using finite elements, the principle of electromagnetic reciprocity is applied.The governing relationship for the Jacobians of the field along strike is obtained bydifferentiating the Helmholtz equation with respect to the conductivity of each regionin the finite element mesh. According to the principle of electromagnetic reciprocity,the roles of sources and receivers are interchangeable. Utilizing reciprocity, the fieldvalues obtained from the original forward problem and for new unit sources imposedat the receivers are then utilized in the calculation of the Jacobian by simplemultiplication and summation with finite-element terms at each rectangle in the mesh.Reciprocity decreases the computational expense of Jacobians to the order of thenumber of stations of interest. After the first inversion, the sensitivities matrix isupdated by the Broyden updating formula, which saves the computational timegreatly.(4) The approach of Reweighted Conjugate Gradient is based on regularizationtheory and adaptive regularized factor. Conjugate Gradient Algorithm is utilized tosearch for the minimum of the parametric functional. The optimal regularizationparameter is updated on subsequent iteration if the misfit does not decrease fastenough. This procedure insures the convergence of the inversion and enhances thespeed of inversion calculation.(5) The simulated results of several synthetic data and imaging of field data inXuwen area are presented by focusing inversion, which are higher resolution thanusing conventional inversion method of MT data. All of these indicate the inversionapproach can describe the blocks structures correctly.