Studies On The Method Of Multi-attribute AVO Analysis & Elastic Parameters Inversion

AVO Analysis, a new technique in extraction of lithology parameters, plays an irreplaceable role in the area of oil-gas exploration. Integrating with techniques of AVO polarization, cross-plotting, fluid replacement model, AI and EI inversion, Lambda-Mu-Rho (LMR) and using the sections of P-wave, S-wave, Gradient, near and far angle stack, we can extract the rock and fluid physical parameters and predict the reservoir lithology and fluid saturation characteristics. Up to now, in the respect of using different lithology models to study the AVO characteristics, the predecessors have already done much work, which is just a beginning of jointed using AVO analysis and seismic inversion to study reservoir property. Especially, there is not a mature theoretical system in jointed using multi-wave AVO to quantitatively depict reservoirs. In order to implement refined parameters inversion and improve the capacity of reservoir prediction with multi-wave AVO, we mainly focus on the two aspects of multi-attribute AVO analysis and elastic parameters inversion to discuss the theory and method of AVO forward & inversion based on the summaries of AVO background, history, state of the art and development trends.The fundament of AVO analysis is the formula of reflection coefficient denoted by different elastic parameters, so the paper firstly discuss the relationship between Zoeppritz equations and the formation parameters, and the information involved in the reflection and transmission amplitude of plane wave in detail. Secondly, referring to the author's former research findings, the qualitative and quantitative comparisons of the approximate reflection coefficients (be published), we present several high-precision reflection coefficients formulae for PP and P-SV wave from different aspects. Also we discuss the precisions of the new formulae using different models, which lay a foundation for the subsequent AVO analysis of multi-attribute & elastic parameters methods. Considering the special significance of AVO attributes and multi-dimension AVO in reservoir prediction, combining the research hotspot and the advancing edge of AVO technique, we systematically analyze the attribute characters of the new approximations aforementioned, and establish the multi-attribute & elastic parameters AVO analysis scheme for PP and P-SV wave respectively. ⑴We thoroughly discuss the physical meaning, the relationship between AVO attributes, background trends, characters of AVO cross-plotting and the implications in parameters inversion for the new formula represented by Lame contrast and shear modulus, and present an elastic parameter AVO analysis method. By analyzing physical signification of newly compound parameters, we find that the two parameters reflect the impedance change of P-wave and S-wave in the seismic attribute analysis while presenting the change of liquid factor or Poisson's ratio in petrophysics. Therefore, by using AVO background trends and cross-plotting characteristics of these two parameters, not only we can obtain the ratio of compressional-to-shear-wave velocity (), but also magnify the weak AVO abnormality. Furthermore, the ratio of has been hidden in the two new parameters aforementioned, which provides a guarantee for quantitative AVO parameters inversion. ⑵To synthetically use multi-attribute AVO to extract more lithological information, we analyze the attribute characteristics of the new reflection coefficient formula denoted by different angle and AVO multi-attribute, and of higher-order approximate formula. Meanwhile we discuss the methods of constructing multi-attribute seismic section and multi-dimensional AVO cross-plotting, and develop the method of AVO multi-attribute analysis by discussing the characteristics of multi-attribute AVO cross-plotting under different models. ⑶There are many factors to be considered in the elastic impedance inversion: First, P-wave and S-wave velocity have to be constants on the whole profile. Second, the seismic dataset have to be superposition of the same angle, which...