The results of modified optimization algorithm for nonlinear AVAZ inversion are presented in the paper. Data of reflected compressional waves in anisotropic media is used. The algorithm based on exact formulas for reflection coefficients applied to processing and interpretation of 3D full-azimuth seismic data. Seismic survey was performed at one of the fields in the Orenburg area. The modified algorithm is differentiated on three stages. At the first stage, second-order Ruger approximation is used to estimate the symmetry axis direction of the anisotropic HTI medium. Using Ruger approximation allows to reduce the number of estimated parameters on the next stages, on which computationally expensive exact formulas are used. At the second stage, shear wave velocity is estimated from data in the isotropy plane. Anisotropy parameters are estimated at the last stage from reflection coefficient data for all offsets and azimuth. Differentiation on different stages allows to increase robustness of optimization, increase accuracy of estimated anisotropy parameters, and decrease computational time costs. The modified algorithm is compared with traditional approach with linearization of reflection coefficients by Ruger approximation. Results are presented in the form of vector maps of anisotropic parameters distribution: parameter for Ruge approxi mation Bani and Thomsen parameter γ for exact formulas. From comparison analysis it could be concluded that AVAZ inversion algorithm based on exact formulas provides higher resolution of parameters distribution. The areas with increased anisotropy highlighted on the vector maps could be interpreted as areas with more ordered reservoir fracturing. Comparison with attribute analysis shows correspondence of results. Spotting areas with increased fracturing from higher anisotropy in combination with other dynamic attributes allows to identify areas with better filtration characteristics of the reservoir, which are more favorable for drilling.
