Fast geometric attribute-constrained prestack seismic inversion for horizontal transverse isotropic media

Abstract Horizontal transverse isotropic (HTI) media, the simplest model of azimuthal anisotropy, are characterized by a parallel arrangement of vertical fractures. HTI media have been extensively studied in the exploration of hydrocarbon reservoirs with vertical or high-angle fractures. However, there are often multiple solutions for fracture prediction based on HTI, owing to insufficient prior fracture information and differences in parameter contributions. Accordingly, we have developed a fast geometric attribute-constrained prestack seismic inversion (FGCI) method based on two geometric attributes: dip and curvature. First, based on numerous variables causally related to fracture density, including fracture-inferring material properties, curvature attributes, and their combined parameters, a curvature-constrained kernel function is constructed using a linear regression model. This function accounts for two types of fractures, structural and background fractures, providing reasonable prior information for fracture prediction and exhibiting a high correlation with fracture density. Second, the expression for the geometric attribute-constrained operators is derived by embedding the curvature-constrained kernel function into the multitrace inversion framework, thereby improving the universality of the curvature for quantitatively characterizing the fracture density. Finally, a computationally well-behaved objective function was established and written in Lagrangian form. Considering multiple constraints, adaptive penalty factors have been developed to balance the residuals of the primary and dual problems, thereby ensuring computational efficiency and stability. Numerical experiments and field data applications demonstrate that this method exhibits strong stability, spatial fidelity, and resolution. Thus, FCGI can provide a stable, efficient, practical, and comprehensive solution for predicting fractures in complex structures.