| As the leading energy in China,coal has played an important role in social and economic development for a long time.With the depletion of shallow resources,the mining depth increases.Water inrush shows gradually the characteristics of complex inducement,non-linear,sudden and difficult to predict in coal mining.Especially for water inrush disasters from the floor with confined water,there are various difficulties in monitoring,early warning,and prevention and control.Therefore,it is of great significance to study and explore the sudden change precursor law of physical and mechanical parameters of historical water disasters,to obtain and use the multi-source information such as rock fracture and hydrological sudden change generated by stope response in the mining process,and to realize the intelligent early warning of water disasters.For the realization of intelligent decision-making,an early warning technology with monitoring of water inrush risk in stope was proposed considering the static prediction and evaluation before mining,microseism(MS)real-time monitoring of floor failure and dynamic prediction in mining.The proposed technology was applied and verified in the experimental engineering to realize the effective prevention and control of water disaster.(1)Combined with the response mechanism of floor damage and failure from seepage and excavation in fissured rock mass,the RFPA numerical software of flow coupled was used to simulate and analyze the initiation and expansion of floor fracture,the tracking and transmission of water pressure,and the evolution process of water-resisting layer into water channel.Shear deformation,damage,and conduction characteristics of floor rock mass was induced by strain incongruity under the superposition effect of mining and seepage pressure.The mechanism of water inrush in typical geological structural areas was revealed.Collapse columns and faults can cause the integrity of coal seam floor,reduce the strength of water-resisting rock,weaken its ability of impedance deformation.Finally,the floor was broken down by the lower confined layer to induce water inrush disaster.(2)Based on the source mechanism of MS monitoring of rock fracturing,a scientific concept of "connectivity" was proposed.The continuous calculation method of the regional "connectivity" of the floor strata and the MS identification criterion of the conduction of the deep failure in the mining process were given.When the connectivity is greater than 1,it can be considered that the two micro-fractures are connected to form a channel,and if less than 1,the two ones will not connect.A lager probability that the mining fissure in the floor will penetrate the water resisting stratum will occur with the greater connectivity,and more likely form a water diversion channel.Through the connectivity calculation and the identification of connection path in floor based on MS monitoring,the dynamic characterization and conduction early warning of water diversion channel can be realized.(3)For the three elements of floor water inrush: water source,water volume,and water diversion channel,based on the coal mining of Group A in Huainan mining area,the monitoring systems of high-precision MS and underground hydrology were built.The coupling analysis various geological factors from two technologies was preliminarily realized.Through the integration analysis of monitoring data,it was concluded that,the regular spatiotemporal and aggregated MS events were induced in mining,and consistent with the distribution of "lower three zones" and the stope weighting.Seasonal variation of hydrological information was accompanied,such as water temperature and PH value.(4)According to the monitoring results of the Internet of things,a new method and technology for early warning of water inrush in coal floor were proposed,including the MS monitoring of floor damage,hydrological monitoring of underground water gushing points(water pressure,temperature,and quality,etc.),water level monitoring of ground observation holes,and intelligent real-time early warning from big data.The Depth Neural Network(DNN)was used to construct the static prediction and evaluation model before mining,and the dynamic prediction and early warning model during mining.The water inrush probability predicted by the trained DNN model was basically consistent with the measured data,and the deviation was kept in the range of ±0.2.The neural network model of Long Short-Term Memory(LSTM)had good convergence in both training set and verification set,and was stable around 17.25%.The predicted results from LSTM were basically consistent with the field measurements.This study realizes the omni-directional real-time monitoring,classification and dynamic intelligent early warning of water inrush from coal floor,and can provide some guidance for water disaster control and intelligent mining.Figure [89] Table [13] Reference [160]... |
PDF Full Text Request