Jont Inversion Of Marine Controlled-source Electromagnetic And Seismic Data Using The Structural Constraints

Joint inversion of different geophysical data collected in the same region can make the inverted image closer to the actual geological settings. It can also overcome the limitations of separate inversion and reduce the impacts of non-uniqueness of geophysical inversion. The seismic method provides high-resolution images of structure and stratigraphy, determining fluid properties such as oil-water interface using seismic data alone can be difficult or, in some cases, impossible. The sensitivity of resistivity to hydrocarbon saturation and fluid properties such as temperature and salinity is well known, and widely exploited in well log analysis and interpretatioa The resistivity attribute derived from marine controlled-source electromagnetic (CSEM) data, if properly interpreted, can provide information which when integrated with seismic data allows fluid properties to be determined with more certainty.This thesis mainly discusses the modeling and inversion techniques of marine CSEM data and seismic data. The modeling of acoustic wave equation is used as a forward driver for full-waveform inversion (FWI). Then we study the join inversion algorithm of marine CSEM and seismic data. The main structure of this thesis is described as follows.(1) Derivation of the semi-analytical formulas of electromagnetic fields in layered isotropic mediaFrom Schelkcunoff potentials, we derive the semi-analytical formulas of electromagnetic fields excited by electric dipole source in layered isotropic media. The electromagnetic field excited by a finite source is computed by the Gauss Quadrature according to the Principle of Electromagnetic Field Superposition. We use the azimuth angle and dip angle to divide the arbitrariely orientated finite source into three equivalent parts along the coordinate axes, respectively. Then, the electromagnetic fields can be obtained by the superposition of electromagnetic fields the three decomposed parts of equivalent source.(2) Marine CSEM 1D inversionFrom above theory of marine CSEM one-dimensional (ID) forward modeling, we derive the semi-analytical formulas of the electromagnetic fields with respect to the earth layer resistivity. Then, we realize the 1D inversion of marine CSEM data by Gauss-Newton method and BFGS (Broyden, Fletcher, Goldfarb, and Shanno) method.Furthermore, we derive the semi-analytical formulas of the electromagnetic fields with respect to the transmitter navigation parameters (including the azimuth, dip, and horizontal location). We decompose the Jacobian matrix by SVD analysis and an eigenparameter analysis shows that that seafloor resistivities and transmitter navigation parameters can be independently resolved. Then, we present a joint inversion method of the transmitter navigation and the seafloor resistivity for frequency domain marine CSEM data.(3) Marine CSEM 2D/3D forward modeling using the staggered-grid finite difference methodWe perform three-dimensional (3D) marine CSEM forward modeling by finite difference method. A single second-order equation of electric field is discreted by staggered-grid finite-difference method, and the electric dipole source is simulated by the pseudo-delta functioa By performing Fourier transform to the strike direction, we realize two-dimensional (2D) marine CSEM forward modeling by staggered-grid finite-difference method using a pseudo-delta function approach.(4) Finite-difference modeling and inversion of acoustic waveWe realize the frequency-space domain 1D/2D modeling of acoustic wave by PML absorbing boundaries. Then, we perform 1D/2D full-waveform inversioa The modified Gauss-Newton method with the diagonal Hessian is used to keep the inversion stable and being of fast convergence. A multi-frequency data-weighting scheme prevents the high-frequency data from dominating the inversion process.(5) Joint inversion of marine CSEM and seismic dataWe realize the joint inversion of marine CSEM and seismic data by using the structural coupling functionals. The JTV (Joint Total Variation) structural coupling functional is used for the 1D joint inversion case and cross-gradient and JTV structural coupling functionals for the 2D case. For separate marine CSEM inversion, the Gauss-Newton scheme is used for fast convergence and accurate resistivity imaging. For separate seismic inversion, the inversion approach is still Gauss-Newton method but a modification is applied for keep the inversion algorithm stable and being of fast convergence. Synthetic tests show that joint inversion of electromagnetic and seismic data is an effective way to improve the reliability of reservoir evaluation. The resistivity image is improved by seismic structural constraint and one can easily identifies the boundary between hydrocarbons and water form the resistivity image.