A new three-dimensional mesh modeling method for geological bodies containing complex faults in geological engineering

Three-dimensional geological modeling and visualization hold considerable significance for both application and research across various fields, including geosience, marine geology, environmental protection, mining engineering, and geological engineering. However, the technique and approach are exposed to critical challenges for complex geological bodies, including heightened structural complexity, extensive manual intervention, and diminished accuracy of discontinuity surfaces owing to the presence of faults, particularly conflicting discontinuity surfaces generated by reverse faults. In this paper, a novel technical workflow and a set of three-dimensional structure modeling methods for geological bodies containing complex faults were designed. A model stitching strategy based on the ear clipping algorithm was proposed to incorporate fine fault models into the modified original model. This strategy transforms the integration of complex three-dimensional models into a triangulation problem for simple two-dimensional polygons, ensuring topologically consistent and computationally efficient merging. Simultaneously, a contour line segments sorting algorithm based on a vertex sequence was proposed to accelerate the extraction of boundary contour line segments. We used a two-stage detection strategy to accelerate locating overlap regions. A multi-layer triangulated irregular network-3DT model was constructed through continuous stratum modeling, fine fault modeling, overlap detection based on the two-dimensional projection topological relationship, extraction of ordered contour line segments of model boundaries, and model reconstruction based on the ear clipping. An underground modeling experiment revealed that these modeling methods and technical systems can accurately depict the geometric form around complex geological faults. A quantitative and qualitative comparison between the proposed modeling method and various modeling methods demonstrates that our method is superior in flexibility, robustness, and mesh quality.