Failure Law Of Floor In Deep Mining And Prediction Of Floor Failure Depth Based On Weka Platform

With the increase of the coal seam mining depth,the problem of mine disasters is more and more serious,especially the problem of floor water invasion,it threatens to safety production in coal mines seriously.The study of the floor failure law in deep mining,especially the depth of destroyed floor,is of great significance for understanding and preventing the floor water invasion.Therefore,this paper studies the floor failure law in deep mining by means of theoretical analysis,numerical simulation and similar materials.The floor stress,the displacement distribution characteristics and the development law of the plastic zone are analyzed.Be based on the Weka platform,a prediction system of the floor failure depth in deep mining is established,and the following research results are obtained:(1)The peak values of vertical stress and shear stress gradually increase with the working face advanced,the peak values of vertical stress and the shear stress are 76.8MPa and 26.7MPa respectively when "the length is equal to the length of working face advancing".The depth of destroyed floor increases with the advancing of the working face,and the depth of destroyed floor is 30m after "the length is equal to the length of working face advancing".As working face continues to advance,the maximum depth of destroyed floor no longer increases and is basically stable.The deformation of the floor rock is inverse ratio to normal distance of the bottom plate,and the farther distance from the bottom plate is,the smaller the deformation of the bottom plate is.(2)According to the mining and geological conditions of a certain coal mine,the similar material simulation experiment method is used to simulate the movement of floor strata in deep mining.The analysis shows that,the stress value is decreasing with the increase of the floor depth,and the stress propagation has hysteresis effect.With the increase of the depth,the lag time is longer.(3)The mining depth,working face length,mining thickness,mining method and fault on the depth of destroyed floor are studied by the FLAC30 numerical experiment.With the increase of mining depth,mining thickness and working face length,the maximum failure depth increases.The backfill mining(filling rate 75%)is 25%lower than the caving mining.The maximum failure depth is affected by faults,and the maximum failure depth is increased by 33%compared with that without faults.(4)The national geological data about the destroyed floor depth is collected and arranged,mining depth,mining thickness,coal seam dip angle,fault and floor resistance damage potential are treated the influencing factors,learning and training the samples through Weka platform and analyzing the results,it is concluded that the optimal model is artificial neural network model.Be based on the Java language,the prediction system of destroyed floor depth is developed and applied.The results show that the Weka prediction model is more accurate,applicable to mining conditions of the mining depth in 110m～550m,coal seam dip angle of 0°～15°,working face length in 30m～135m.