Seismic-petrophysical inversion based on linear Bayesian theory and modified canonical correlation analysis
Petrophysical inversion,as an intuitive quantitative interpretation tool,is an important topic in reservoir characterization,where Rock Physics Models(RPMs)link petrophysical and elastic parameters.Traditional petrophysical inversion usually requires calibrating RPMs at well location.In addition,it is difficult for regular RPMs to accurately describe the relationships between petrophysical and elastic parameters in a complex geological environment,which further affects the accuracy of petrophysical inversion.To solve these problems,a modified canonical correlation analysis method called BP-CCA is introduced to construct statistical petrophysical relationships between petrophysical and elastic parameters,so as to obtain the subsurface spatial distribution of petrophysical parameters.Moreover,the linear Bayesian theory is adopted to invert elastic parameters,such as the P-and S-wave velocity and density,from prestack seismic data,which has higher inversion accuracy and computational efficiency.In our work,both synthetic data experiments and practical data applications are carried out using the proposed method,which prove that the proposed method provides higher reliability and more precise characterization of petrophysical parameters than traditional methods.
NETL
NSTL
万方数据
维普
