Two-dimensional inversion of CSAMT tensor data

摘要

CSAMT plays an important role in the field of oil and gas exploration, mine survey , engineering and environment detection and so on[1]. Scalar survey is usually used in CSAMT exploration. However, the research of MT inversion[2] has shown that the inversion results with tensor data are better than those with scalar data and tensor data should be used in field data inversion. A 2D modeling and inversion algorithm has been developed for CSAMT tensor data to promote the application of CSAMT. The secondary electric and magnetic fields generated from 2D model with 3D source in the wavenumber domain satisfy the Maxwell's equations. After discretization with finite element method, a forward equation system is assembled. The secondary electric and magnetic fields in the wavenumber domain by solving the forward equation. After inverse Fourier transform, the total electric and magnetic fields in the space domain can be obtain by adding the primary fields calculated by 1D modeling. Then, the tensor apparent resistivity and phase responses on the surface can be computed. OCCAM method is used in the inversion of tensor CSAMT data. The Jacobian matrix is calculated by employing the adjoint-equation approach. The test model is shown in Fig. 1. Two electric dipole transmitters inclined to the survey line (x=0m) at an angle of 45 degree are located at (0, 0, 0) m. A 2D conductive prism ( ρ=10Ωm) of size 280×240m along the x- and z-coordinates, respectively. The target with strike in y direction is embedded into a homogeneous and resistive background (ρ=100Ωm) and its upper boundary lies at a depth of z=40m. The tensor apparent resistivity and phase data, ρ_(sxy) ,Φ_( xy) ,ρ_( syx) and Φ_(yx) , at 10 frequencies (1000-1Hz) were generated using our 2D modeling program. Two percent normally distributed Gaussian noise was added to the data. These synthetic data were inverted using our 2D inversion program. The inversion result shown in Fig. 2 demonstrated the validity and stability of the inversion algorithm.