| The full waveform inversion(FWI)method is a nonlinear inversion method based on the measured seismic data and simulated seismic data using the optimization method for seismic imaging.Since the full waveform inversion method makes full use of the wave field information such as amplitude,travel time and phase,it can not only provide a good elastic parameter model for reverse time migration,but also help to oil and gas reservoirs by imaging the elastic parameters.Since the elastic wave equation can describe the propagation characteristics of seismic waves more precisely,the full waveform inversion of elastic wave has always been the focus of theoretical research.The finite difference contrast source inversion method is an elastic wave full waveform inversion method based on the inverse scattering theory.This method not only can effectively reconstruct the elastic parameters,but also can improve the efficiency of the inversion.The finite difference contrast source inversion(FDCSI)method is a nonlinear inversion method under the framework of inverse scattering theory.Unlike the conventional FWI method to directly reconstruct the elastic parameters of the formation,the FDCSI method estimates the elastic parameters by reconstructing the contrast source(virtual scattering source)and contrast.We apply the FDCSI method to the elastic wave equation for multi-parameter inversion to reconstruct the P-wave and S-wave velocities.This paper mainly carried out the following work: First,the conventional elastic wave forward modeling and contrast source forward modeling.We use the optimized 25-points finite difference method to construct the frequency domain difference operator and obtain the simulated scattered wave data.By sparsely storing the operator,the frequency domain forward modeling is more convenient,taking up less memory and more suitable for multi-source excitation test.Secondly,finite difference contrast source inversion.Since the difference operator is only related to the background medium and the given frequency,the construction of the difference operator and the matrix decomposition only need to be performance once in each iteration,which greatly improves the inversion efficiency,and also shows that this method has great potential of 3D waveform inversion.Finally,we demonstrate the effectiveness of the proposed method in a simple model and the ability to perform high-precision imaging in complex models through numerical simulations of simple models and Marmousi models. |
