A RBFNN-based method for the prediction of the developed height of a water-conductive fractured zone for fully mechanized mining with sublevel caving |
| |
Authors: | Qiang Wu Jianjun Shen Weitao Liu Yang Wang |
| |
Institution: | 1.National Engineering Research Center of Coal Mine Water Hazard Controlling,China University of Mining & Technology (Beijing),Beijing,China;2.College of Mining and Safety Engineering,Shandong University of Science and Technology,Qingdao,China |
| |
Abstract: | The movement and failure of overlying strata induced by underground coal mining cause “three zones,” including the caving zone, the water-conductive fractured zone, and the sagging zone from the bottom up. For knowledge about the height of the water-conductive fractured zone, there has been no empirical or theoretical formulae for thick coal seam using fully mechanized longwall mining with sublevel caving. This paper presents a methodology of determining the height of the water-conductive fractured zone based on the radial basis function neural networks (RBFNN) model in MATLAB software. Before modeling, the relationship between the height of the water-conductive fractured zone and mining thickness, the lithologic character of the overburden and its composite structures, and workface parameters was studied. After that, 32 and 7 measured data were used as training and testing samples, respectively. It has been found that the average relative error is 6% and the maximum relative error is 10% for 7 test samples by comparing actual results with predicted results. The model was applied to the no. 31503 workface in the Gaozhuang coal mine for safety evaluation. The predicted value is 59.6 m, and the measured value is 55.9 m. The RBF-based model shows much better performance than empirical formulae in the Regulations for Coal Mining and Coal Pillar Design under Buildings, Water-bodies, Railways and Main Shafts for the prediction of the height of the water-conductive fractured zone for fully mechanized mining with sublevel caving. |
| |
Keywords: | |
本文献已被 SpringerLink 等数据库收录! |
|