Azimuthal AVO inversion by simulated annealing

David Cho, Gary F. Margrave

Fractures influence the permeability pathways and mechanical properties associated with a rock mass and therefore, are a crucial aspect in the characterization of the subsurface. In this study we develop an azimuthal AVO inversion algorithm using a simulated annealing optimization technique. The parameterization of the problem is in terms of an isotropic background with the inclusion of fractures through an addition of excess compliances to the medium. Preliminary inversion results demonstrate a reasonable estimate of the model parameters in addition to an excellent match between the data and synthetic data. Associated errors in the estimated model parameters are attributed to variable sensitivities of the model parameters to the objective function. Future work will attempt to address these issues through different parameterizations of the problem and additional constraints in the objective function.