The Law Of Surface Movement And Deformation And The Construction Of Regional Prediction Model For Mining Under Thick Loose Layers

The increasing intensity of coal resources leads to the further severe problem of"three under" mining pressure.Huainan is located in the hinterland of East China,and it contains abundant coal resources,but a large number of coal seams are buried under thick loose layer.The surface subsidence caused by mining under thick loose layer will show many characteristics,such as the maximum subsidence coefficient is greater than 1,and the boundary convergence is slow.The above characteristics cause the limited accuracy of mining subsidence prediction under thick loose layer,especially the prediction outside the inflection point of subsidence basin,and the relative error often reaches more than 60%.Therefore,in order to better guide the safety of production,it is very important to establish a model adapt to the mining subsidence prediction model under thick loose layer,and is easy to realize and parameters can be easily obtained,which is very important to protect the ecological environment,sustainable development of coal resources and prediction of damage level of surface buildings(structures)in the area.Taking Huainan mining area as the research object,based on the analysis of measured data,the surface movement law of mining subsidence under thick loose layer were studied in this paper,and a prediction model with strong adaptability to mining subsidence under thick loose layer were constructed,and the model to solute the initial value of the parameters of the prediction model of Huainan mining area were also established in this paper.(1)Through the detailed analysis of the dynamic and static law of mining movement and deformation of single workface and the comprehensive analysis of mining subsidence parameters collected in Huainan mining area,the special law of mining subsidence under thick loose layer is summarized;The solution model between mining subsidence parameters and geological condition parameters is constructed by grey correlation method combined with successive regression and fusion of independent variables.Two precision indexes of fitting mean square error and generalization performance of the model verify that the solution result of the model is reliable and has practical prediction value;then the combined model for layered prediction of rock and soil is established.The model regards the overlying strata and loose strata of coal seam as two different media,then the probability integral model is used to predict the movement and deformation caused by them respectively,and then the two models are weighted and combined according to the dynamic linear weight.By adjusting the parameters of the model,the convergence rate of the prediction subsidence basin can be controlled.(2)In order to accurately calculate the parameters of the probability integral model and the combined model for layered prediction of rock and soil,and solve the problem that the traditional optimization algorithm has a strong dependence on the initial value of parameters when solving the nonlinear model parameters,which limits the parameter solution of the new model.So the cuckoo search algorithm were introduced in this paper with its local search ability enhanced.It solves the problem that the local search ability of the original algorithm is not strong.Then the simulation results show that the algorithm has good resistance to the problems of random error,coarse error and missing observation point in observation data.The accuracy of the algorithm is proved to be reliable by comparing with the inversion effect of genetic algorithm and total least squares robust algorithm.(3)Seven mining faces under thick unconsolidated strata in Huainan mining area were selected and fitted by the probability integral model and the combined model for layered prediction of rock and soil.The results show that the new model can ensure the high prediction accuracy of probability integral model in the middle part of the subsidence basin,and the fitting accuracy of the edge part of the subsidence basin can be improved from more than 60%to less than 20%.The calculation model between the parameters of the combined model for layered prediction of rock and soil and the probability integral model is constructed.Combined with the calculation model of probability integral model about the geological condition parameters,the calculation model of initial value of the combined model for layered prediction of rock and soil parameters is indirectly obtained.(4)Based on the theoretical research,a software for mining subsidence prediction in Huainan mining area is developed successfully,which integrates the calculation of the mobile deformation value,the parameter solution and the prediction of the moving deformation value.Figure 80 Table 41 Reference 199...