Three-dimensional Forward Modeling And Inversion Of MCSEM Data In Anisotropic Medium

In the past ten years,the marine controlled sources electromagnetic method(MCSEM)is widely used in marine geological structures and oil and gas resource exploration due to its low cost and high detection accuracy.Nowadays,the advances in the development of instruments have greatly improved the data quality and acquisition efficiency such that 3-D dense survey layouts are becoming routine for acquiring large volumes of data.A comprehensive data coverage can enhance the imaging resolution of 3-D structures and thus reduce the ambiguity in data analysis.However,as the difficulty of exploration increases,this also creates the need for developing sophisticated interpretation tools such as 3-D inversion algorithms to treat large data sets.Obviously,the development of marine controlled sources electromagnetic method data processing and analysis technology under complex medium conditions is relatively lagging.At present,most of the MCSEM forward modeling at home and abroad is based on the single isotropic property of the dielectric and ignoring the influence of the anisotropy of the medium and the physical parameters of the rock.At the same time,the stability and efficiency of 3D MCSEM inversion need to be further improved.Although assuming isotropy can be sufficient in a number of situations,the presence of electric anisotropy in the earth's crust,due to thin layer interbedding or grain alignments in the sediments can significantly alter the response measured at the seafloor Failure to properly treat anisotropy can produce misleading and sometimes uninterpretable results.As a consequence,accounting for anisotropic effects becomes a necessary challenge to explaining results that isotropic models cannot.Moreover,most of MCSEM inversion algorithm system use a regular mesh for inversion,obviously,it is insufficient to reflective of the geometric characteristics of complex underground geological structures.In order to solve the problems mentioned above,in this paper,the three-dimensional numerical simulation of marine controlled sources electromagnetic methods in isotropic and anisotropic medium is carried out.And the effects of conductivity anisotropy and rock structural parameters on the electromagnetic response of MCSEM are analyzed in detail.In addition,the three-dimensional inversion of MCSEM data under the condition of conductivity anisotropy is carried out by using unstructured mesh and the effect of conductivity anisotropy and terrain on inversion is discussed.Firstly,this paper deduced the MCSEM control equation of the total and secondary field formulations in detail.Then,we use an edge based finite element method to solve the MCSEM forward modeling problem in anisotropic medium.We adopt the unstructured mesh to Discretization the model domain and it is conducive to the construction of complex geological models.The open source parallel direct solver MUMPS is used to solve the equations of the finite element method,so the efficiency of solving the equations can be effective improved.We first introduce an incremental model to study the effects of some petro physical parameters on conductivity and electromagnetic response in the isotropic medium.What is more,the electromagnetic response characteristics of MCSEM under the condition of spindle anisotropic medium are also analyzed.Then,based on the secondary governing equations satisfied by the coulomb's vector-scalar potential,the node based finite element method based on unstructured mesh is used to carry out the 3D MCSEM forward modeling in arbitrary anisotropy of medium.The IDR iterative algorithm based on incomplete LU decomposition technique is applied to solve the finite element method linear equations system,which improves the efficiency of solving equations.The moving average least method(MLSM)is used to derive the vector and scalar potentials,which effectively improves the calculation accuracy.The influence of arbitrary anisotropy of conductivity on the MCSEM response is analyzed in detailed.In the MCSEM inversion study,the inexact Gauss-Newton method is used to realize the 3D inversion of MCSEM data in VTI conductivity anisotropy medium.We use a quasi-forward modeling method to calculate the product of the Jacobian and its transpose matrix and a vector,which can effectively avoid to explicit calculate and storage of the Jacobian matrix,thus saving computation time and storage space.We adopt an unstructured mesh in the forward modeling and inversion processes.Therefore,it is possible to simulate arbitrary undulating terrain and complex geological structures and we adopt a strategy of forward modeling and inversion mesh separation in the inversion process.We calculate the transfer between the two sets of mesh parameters by calculating the coincident volume parts of the two sets of meshes.So that we can effectively reduce the uncertainty of inversion without losing the accuracy of forward calculation.In addition,in order to reduce the calculation time of the inversion,we use parallelization for the key parts of the forward modeling and inversion.We designed several theoretical models to calculate the synthetic MCSEM data in anisotropy medium.The three-dimensional results of MCSEM synthesis data show that the developed 3D MCSEM inversion algorithm in VTI conductivity anisotropy is reliable and effective.