Exploration of 3D Coal Seam Geological Modeling Visualization and Gas Content Prediction Technology Based on Borehole Data

The geological structure of coal mines and the precise prediction of coal seam gas content are key factors in creating the transparent working face, and they also represent an important aspect of intelligent coal mining. The traditional technology of coal seam geological construction and gas content prediction is not advanced. This paper presents a methodology for 3D implicit geological modeling and visualization using Gempy and PyVista libraries, as well as gas prediction and distribution based on the Scikit-learn library, all of which are underpinned by machine learning techniques. Under this method, the geological modeling of coal seam was converted to the kriging interpolation algorithm based on machine learning of coal seam thickness data. The problem of coal seam gas content is converted into a regression prediction problem of coal seam characteristic values and gas content target values based on machine learning. The pykrige package under Python is used to interpolate the obtained coal seam thickness. Based on the linear regression prediction model, loss function and other prediction methods and algorithms, the accurate prediction of coal seam gas content based on borehole data is realized. Under the above various operations, a 3D geological model of the mine and the gas content distribution map of the coal seam are finally obtained. Compared to actual borehole data and gas geological maps, this method offers high precision and enhanced efficiency.