Abstract: Well constraints, as one of the important constraints in gravity data inversion, can effectively suppress the inherent non-uniqueness of inversion. However, existing well-constraint approaches still suffer from the problem that overly coarse mesh discretization introduces constraint range errors, while overly fine discretization leads to a significant decrease in computational efficiency. Moreover, well information tends to interfere with the inversion results in non-well regions. To address these issues, this paper proposes a well-constrained gravity data inversion method based on an adaptive mesh refinement strategy, and simultaneously constructs an improved well-constraint weight that is synergistic with the adaptive mesh refinement characteristics. The proposed method employs local refinement around the wells to improve constraint accuracy, while retaining coarse discretization in the background mesh to control computational cost, thereby achieving a balance between accuracy and efficiency. Furthermore, the well-constraint weight is utilized to limit the effective range of constraints, reducing the interference of well information on the inversion in non-well regions and enabling precise control of well constraints. The results of synthetic model experiments demonstrate that the proposed method can comprehensively enhance the resolution and computational efficiency of inversion results. Compared with coarse-mesh inversion, the model relative error is reduced by approximately 20% without significantly compromising computational efficiency; compared with fine-mesh inversion, the computational efficiency is improved by about 50% while maintaining comparable inversion accuracy. The method also enhances the delineation of geological boundaries, and the improved well-constraint strategy effectively optimizes the focusing of inverted solutions and improves the inversion accuracy in non-well regions. Finally, the proposed method is applied to field data from the San Nicolas sulfide deposit, and the inversion results confirm the effectiveness and superiority of the method.

Key words: gravity inversion, adaptive mesh, borehole constraint, modified borehole constraint weight