Theoretical Study On Real-time Monitoring And Early Warning Of Mine Water Inrush

This article focuses on the scientific problem of real-time water inrush monitoring and early warning indicator system and early warning system construction.Based on the actual survey and monitoring index data of Hengyuan Coal Mine,a water inrush warning index system was established,and combined with the mining area conditions and numerical simulation,the identification criteria and graded warning criteria for each index abnormality were proposed.Secondly,the risk theory and expert survey methods were combined to establish a multi-indicator early-warning model for the risk of water inrush from aquifer of limestone of Taiyuan Formation.Finally,a BP neural network water inrush warning model was constructed with the help of matlab.The thesis mainly achieved the following results:?1?An early-warning index system for the real-time water inrush early-warning system was established in the mining face.The index system includes water level,mine water inflow,water temperature,hardness,Ca²⁺,Na⁺,TDS,p H,stress,displacement,and number of microseismic events.According to the current monitoring indicators of coal mines and related equipment,the real-time measurability of each indicator was determined,and the effectiveness of each indicator was determined based on the data of each indicator in the case of water inrush in the mining area and adjacent mining areas.Combining measurability and effectiveness,early warning indicators are rated and selected 6 indicators including aquifer water level,mine water inflow,water temperature,Na⁺,Ca²⁺,and TDS in Heng Yuan coalmine.?2?Risk assessment of water inrush from seam floor.At the?632 and?633working faces in Hengyuan coalmine,with the analysis of the hydrogeological conditions and the calculation of the unit water inflow and the water inrush coefficient of the coal floor,the danger of water inrush from the floor has been evaluated,whose result shows that it is necessary to build a mine water inrush monitoring and early warning system.Finally,a conceptual model of hydrogeology was established for numerical simulation.The numerical simulation determined the evolution law of the flow field in the study area and laid the foundation for the determination of the water level threshold.?3?For each early warning indicator,single indicator abnormality discrimination criteria and graded early warning standards are established.According to the data recorded by the indicators of the mining area in the past 10 years and their normal changes,the abnormal warning thresholds of the indicators have been proposed.In the guidance of the relevant standards of the"Details for the Prevention and Control of Water in Coal Mines",and in view of the changes in the indicators in large,medium and small water inrush cases,a 4-level grading warning method for each indicator is proposed,and a quantitative warning criterion description is given in accordance with the actual situation of Hengyuan coalmine.?4?A multi-indicator early warning model for water inrush risk was constructed and the error of the early warning system was analyzed.Firstly,combined risk theory and the weight calculated by AHP based on expert decision and judgment,a linear risk early-warning model was established,whose comprehensive evaluation risk range was divided.Secondly,using BP neural network theory,a 6-5-5-1 BP neural network model is constructed to learn the training and prediction of the collected data.The error is small,and the prediction accuracy rate reaches 81%,which realizes the requirements of a multi-index comprehensive discrimination early warning model.Based on the working faces of Hengyuan Coal Mine?632 and?633,this paper systematically and quantitatively constructed a pre-grey water risk early-warning model consistent with the actual situation in the study area.It provides ideas for the early warning of this mine and its neighboring mines,and provides a reference for the establishment of real-time water inrush warning systems for other coal mines.There are 66 figures,33tables and 115 references in this paper.