The Application Study In Seismic Inversion Based On Optimized Markov Chain Monte Carlo Method

Abundant information of formation lithology,reservoir physical propery and pore fluid can be obtained by the seismic inversion,used for the reservoir prediction and fluid identification.With the increasing complexity level in reservoir exploration,the nonlinear geophysical inversion methods with high presicion and resolution are playing an increasing important role in the geophysical inversion.Markov chain Monte Carlo(MCMC)is a nonlinear probabilistic inversion method i global optimization,avoid the complex problems of marginal integral calculation with the unique structure.Based on the statistical analysis of created Markov chain consistent with the posterior probabilistic density,we can get the statistical characteristics of the complex multidimensional posterior distribution,and fially obtain the optimal solution and its associated uncertainty.The thesis is aimed at the MCMC method,and based on the study of adaptive updata strategy of the proposal distribution,we propose an optimized MCMC method in the Bayesian framework,accelerating the running speed of MCMC and improving the converging effect of Markov chain.O n the basis of isotropic media hypothesis,we propose a method of optimized-MCMC-based EVA/AVA(Elastic impedance Versus Angle,EVA;Amplitude Versus Angle,AVA)nonlinear inversion based on the Bayesian theory.In the Bayesian framework,the reservoir fluid sensitive indicator,that is effective pore-fluid bulk modulus,can be inverted stably based on the EVA inversion method,used for the fluid identification of oil and gas reservoir.Moreover,the nonlinear AVA inversion method based on the optimized MCMC using the exact Zoeppritz equation is not only suitable for reservoirs with strong-contrast interfaces and long-offset ranges but also further reduces the non-uniqueness of the solutions based on the Bayesian framework and the fusion of a priori constraint information such as logging data and seismic data,and greatly improves the stability of the inversion solutions,estimating the uncertainty of the results to assist us in risk assessment of reservoir prediction.To overcome the unstable fracture weaknesses inversion problem,we also propose a method of optimized-MCMC-based EVAZ/AVAZ(Elastic impedance Versus Angle and Azimuth,EVAZ;Amplitude Versus Angle and Azimuth,AVAZ)nonlinear anisotropic inversion based on the anisotropic media hypothesis.Based on the derivated EVAZ equation,we can use the ration of azimuthally anisotropic elastic impedance to eliminate the influence of the isotropic background elastic impedance to the anisotropic disturbance elastic impedance and obtain the normal and tangential weaknesses more stable for reservoir fracture characterization.In addition,based on the derivated AVAZ equation,we can use the azimuthally seismic differences to eliminate the influence of isotropic background reflection to the anisotropic perturbation reflection and obtain the fracture weaknesses to provide an important geophysical support for the exploration and development of the fractured reservoir.