Application of NIPOS-SVM Model for Evaluation of Water Richness of Coal Seam Roof Aquifer—A Case Study of the Xinhu Coal Mine in Huaibei, China

The sandstone roof of coal seams, with its high porosity and developed fissures, serves as a favorable reservoir for groundwater. Predicting and assessing the water-bearing capacity of the sandstone roof in coal seams is crucial for the rational development of coal tunnels, ensuring safe and efficient production in mining areas. This study targets the Cenozoic bottom aquifer of the No. 81 mining area of the Xinhu Coal Mine. By analyzing the geological and hydrogeological conditions of the mining area, it was found that the primary water-bearing strata of the coal seam roof are the Permian sandstone fracture waters. Key factors for evaluating the water richness of the sandstone aquifer were identified as aquifer thickness, aquifer depth, core recovery rate, coal seam dip angle, brittleness–plasticity ratio, and the sand–mud interlayer index. A novel particle swarm optimization algorithm incorporating improved sine chaos mapping (NIPSO) to enhance the support vector machine (SVM), thereby constructing the NIPSO-SVM model, was applied for quantitative evaluation of water richness in the study area. Experimental results indicated that the NIPSO-SVM model has high accuracy and practical engineering application value in predicting water richness, which is significant for ensuring the safe production of coal mines.