Edge reserving pre-stack inversion and its application on sweet spot for shale gas

Economic shale gas production requires hydraulic fracture stimulation and the fracturing effects largely depend on the geo-mechanical brittleness of formation. For this reason, predicting fracture-prone sweet spots is essential for optimizing the development program. The geo-mechanical property of shale can be evaluated by elastic parameters that obtained from three-term AVO inversion. Since pre-stack AVO inversion is an ill-posed problem, the Bayesian framework is employed for its regularization in this paper. Based on normal Bayesian inversion, an edge preserving constraint is introduced on the perturbations of elastic parameters in both vertical and horizontal directions, which could improve the inversion result near geological boundaries and depict the heterogeneity of shale better. The sound anti-noise ability and robustness are fully proved in synthetic data while edge preserving effect is also demonstrated with a 2D model. In practice, this method is also applied on a shale gas play in Sichuan Basin of China and the inverted results show good agreements with well logs. Finally, the brittle sweet spots are discriminated on the basis of elastic property analysis from wells, and an evaluation criterion of gas shale using seismic inverted elastic parameters is established resultantly.