Research On Pre-stack Seismic Inversion For “3+Q”parameters

Reflection and transmission of general plane waves at the boundary between two homogeneous half-space is one of the fundamental problems of seismic wave propagation. The study of reflection coefficient plays an important role in analysis of amplitude variations with offset(AVO), which can be used for direct detection of hydrocarbons. The algebraic complexity of accurate reflection coefficient impedes its usage in practice. Conventional AVO analysis in elastic media is based primarily on approximate solutions of the reflection/transmission problem. In this thesis, a similar approach is used to illuminate the influence of anelasticity on the reflection response in anelastic media. Towards the ambitious destination of inverting anelastic parameters by pre-stack seismic inversion, three aspects of research are conducted.Wave induced fluid flow is the main reason for seismic wave attenuation. Based on White’s layered model, the extended periodic layered model is constructed, which considers the inhomogeneous distribution of both rock frame and fluid. The dispersion and attenuation characteristics of wave propagation in this medium is studied. When there is mesoscale inhomogeneity, the newly constructed model can describe the attention in seismic frequency band. By choosing appropriate length parameters, the extended periodic model can degenerate to White’s layered model.The scattering matrix of anelatic media is constructed using general plane wave functions and the accurate P wave reflection coefficient is given. When the attention degree changes, the characteristics of P wave reflection with incidence angle and frequency change is studied. Assuming small perturbations of anelastic parameters, linearized approximation of P wave reflection is derived. As the quality factor increase to infinite, the approximate coefficient degrades to Aki-Richards equation. We further simplify the reflection coefficient using viscoelasticity for the purpose of seismic inversion. Meanwhile, based on anelastic parameters, the elastic impedance for anelastic media and a new concept called attenuation fluid factor are proposed in this thesis.The approximate P wave reflection formula aforementioned is introduced to construct pre-stack seismic inversion method under the frame of Bayesian theory. Model constraints introduced in the inversion method can not only increase stability but also provide useful low frequency information. The stability and robustness of inversion method is tested using modeling data with added stochastic noise. Applications of the newly constructed inversion method to field data show that absorption parameter related with reservoir fluid can be extracted stably. The absorption parameter have high sensitivity to hydrocarbons in reservoirs and when combined with conventional fluid factor, a more accurate discrimination of reservoir fluid can be achieved.