ESTIMATION OF THOMSEN’S ANISOTROPY PARAMETERS BASED ON AVA RESPONSES

Abstract

In this Master’s thesis, Thomsen’s elastic anisotropy parameters epsilon (ε) and delta (δ) have been estimated as a function of depth at the well site based on AVA responses of real seismic CDP gathers near the well site and synthetic AVA gather modeled for the isotropic case. The estimated anisotropy parameters were used to perform anisotropic synthetic modelling. The AVA response at each of the reflection interface depends on the contrast of Vp, Vs and density as well as on the contrast of anisotropy parameters ε and δ across the reflection interface. Compared with the real seismic data, the modeled synthetic data ignores many propagation effects and is scaled differently. The dataset used in the study is from Colony sand formation, Alberta, Canada. Synthetic AVA gather was generated based on Aki-Richard’s approximation to the Zoeppritz equations for P-P reflection coefficient. Various seismic horizons were picked on CDP gather near the well site as well as on the synthetic gather generated for the isotropic case using Hampson-Russell software package. The Aki-Richard’s coefficients A, B and C of each of picked horizons were used in the study for estimation of the anisotropy parameters. After scaling the amplitudes on the real seismic gather to the synthetic amplitudes, the Aki-Richard’s coefficients B and C for synthetic and real gather were compared and a normalization function was calculated which was then filtered to remove the low-frequency propagation effects. The remaining difference between the low-cut filtered and normalized seismic data (after removing low frequency propagation effects), and the isotropic synthetic data was attributed to the anisotropic part of the reflection coefficients. The resulting distributions of epsilon and delta with depth was compared to the gamma-ray log. The gamma ray and the stimated anisotropic parameters are positively correlated. The estimated anisotropy parameters were the then used for modelling AVA gather for anisotropic case. The modeled AVA gather for isotropic and anisotropic case was compared and a significant difference is observed especially at larger angles. The anisotropic synthetic AVA gather better matches with real seismic gather as compared to isotropic synthetic AVA gather as expected.