Roof water inrush risk assessment based on LDA-RBF and comprehensive weighting method

Abstract

In order to solve the problem of roof water inrush risk during the mining of No. 9 coal seam in Longfeng Coal Mine, an LDA-RBF neural network model for predicting the development height of water-conduction fracture zone was constructed by linear discriminant analysis (LDA), and a CRITIC-AHP comprehensive weighting method was established based on the improved CRITIC evaluation method combined with analytic hierarchy method (AHP) to evaluate the risk of roof cracking and the water richness level of the aquifer in the mining area. Through ArcGis geographic information processing technology, the cracking risk zone and the water-rich zone are superimposed to obtain a comprehensive zoning map of the water inrush risk of the roof of No.9 coal seam. The results show that the LDA-RBF neural network prediction model has a simple structure and higher fitting accuracy, and the development prediction height of the No.9 coal seam water-conducting fracture zone is 50.4 m, which has exceeded the bottom boundary level of most aquifers in the area, indicating that there is a high risk of cracking in most areas. The improved comprehensive weighting method avoids the problem of excessive subjectivity and objectivity of the evaluation results, and the water-rich zoning results are consistent with the actual water inflow of the borehole. Finally, the water inrush danger zone is mainly distributed in strips in the central and northern parts of the mining area, which is the result of the combined effect of strong water richness and high cracking risk of the aquifer in this area, indicating that the above areas should be paid attention to in actual mining.

Keywords

• rbf neural network
• comprehensive weighting method
• arcgis
• height of water conduction fracture zone
• water inrush risk