Yingwang Zhao

Abstract Large language models (LLMs) are increasingly used for information extraction from scientific text, but their reliability in workflows that require structured outputs and executable spatial computation remains uncertain. We evaluate a borehole-report processing pipeline that converts Word documents into structured borehole entities, parses relative location descriptions, resolves survey-control references, and generates start/end coordinates for downstream inspection.

Abstract Water inrush is one of the major mining disasters that may lead to numerous casualties. The development of information techniques makes it possible to monitor the occurrence and evolution of water inrush. Then, locating monitors for water inrush becomes a primary problem. This study presents a method of optimal location of water level sensors by constructing a set covering model.

A new procedure has been developed to improve the velocity field computed by the continuous Galerkin finite element method (CG). It enables extending the postprocessing algorithm proposed by Cordes and Kinzelbach (1992) to three-dimensional (3-D) models by using prismatic patches for saturated groundwater flow. This approach leverages a dual mesh to preserve local mass conservation and provides interpolated velocities based on consistent fluxes.