A Study Of Seismic Prestack AVO Inversion And The Identification Of Gas-Hydrates

Gas hydrate is an ice-like clathrate crystalline compound, composed with water and gas under an environment of high-pressure and low temperature. As the largest untapped new energy on the earth, Gas hydrates have attracted widespread attention. In the field of geophysical exploration, the recognition of gas hydrates and the underlying free gas is the focus of research. Seismic exploration is an effective geophysical method to identify the seabed gas hydrates. The anomaly analysis of the P-wave velocity and Poisson’s ratio of hydrate-bearing sediments is an effective research method; reflection amplitude versus offset (AVO) and incident angle (AVA) of BSR, is also an important method to study the existence of the free gas deposits under BSR.However, in most cases, conventional (stacking and migration) seismic profiles are difficult to identify atypical BSR. It is demonstrated that the P- and S-wave velocity of hydrate bearing sediments have remarkable improvement. Velocity is of great significance in reservoir description. In this dissertation, we use AVO inversion to obtain the P-and S-wave velocity, density of the seabed formation, and it is one of the valid means to identify gas hydrates and free gas.Like most geophysical inversion problems, however, AVO inversion has to face the nonuniqueness and instability of the solution. In essence, the prestack AVO inversion is nonlinear. In some sea areas, the lack of necessary prior information, such as geological background knowledge, logging data and other constraint information, makes the nonuniqueness and instability of the solutions more serious, and increases the uncertainty. Because of the different response between the gas hydrates and oil and gas to the elastic parameters, the presence of hydrate in the sediments causes density decrease while makes the P- and S-wave velocity increase, which violates the basically positive changes in relationships between the velocity and density of the sediments, and thus no longer meeting the generally assumed Gardner empirical formula.Based on the existing identification method of gas hydrate and free gas, this dissertation describes the physical properties and the occurrence models of gas hydrate. Based on AVO forward modeling theory, it investigates the simulation and analysis of different sensitivity of elastic parameters for the primary and transformed reflection coefficients in detail under various parameterizations. We find that, at small incidence angles, the reflection coefficient is not sensitive enough to the change of the S-wave velocity, that is to say, under the same change amplitude of each model parameter, the change of the reflection coefficient caused by the S-wave velocity is minimal, and the effect of P-wave velocity and density is almost identical. While at small angles, this effect is hard to distinguish, this sensitivity difference makes the eigen-values of the three parameters quite different in the process of simultaneous three-parameter inversion, which leads to the different inverted accuracy for the three parameters, especially for the fact that the information of S-wave velocity and density is often not ideal. Also, it gives a detailed derivation and simulation of the AVO responses of the top and bottom interfaces of the hydrate-bearing sediments. This dissertation researches the characteristics of the gradient versus angles based on the Zoeppritz equations, which paves a solid foundation for further research of AVO wide analysis, multi-components and AVO inversion. We perform the AVO inversion simulation in the Bayesian framework with different a priori distributions and quantitatively evaluate the impact of noise on the inversion results, ignoring the Gardner empirical formula. We show the main difficulty and limitations in practice of the Bayesian method. SVM theory is described and the inversion results are improved based on its excellent structural risk minimization. The two AVO inversion methods are compared in the model simulation. Finally, they are applied to the Shenhu area, South China Sea, which has demonstrated its feasibility and effectiveness for the identification of gas hydrates, combined with BSR and other seismic attributes.