Study On Surface-subsidence Dynamic Prediction Model In Mining Area

With the rapid development of China's economy,human social demand for resources is also increasing,which intensifies exploitation of underground resources,such as coal,natural gas,oil and other natural resources.After the lower coal seam is extracted,the original stress balance state between the rock walls around the goaf changes,which makes the surrounding rock strata and the surface deform and move until a new stress balance is reached.In this process,the buildings above the goaf may be damaged and collapsed due to surface deformation,thus causing the loss of life and property of citizens.Therefore,the study of the relationship between the surface deformation and mining time,which has an important guiding significance in the prevention and control of surface deformation in goaf.On the basis of the existing research,this paper explores the problems existing in the dynamic prediction of surface subsidence in mining areas,mainly including the following aspects:Firstly,analyze the existing dynamic prediction time function model and summarize the characteristics of parameters in different models.Second,analyzed the reasons why the Knothe function model and the segmented Knothe function model are inconsistent with the law of ground subsidence.research the physical meaning and the geometric meaning of the time point(?)and the subsidence value(_vw)at the maximum ground subsidence speed,the influence coefficient(c)of the dynamic prediction time of mining subsidence.Thirdly,the theory of variogram is studied and analyzed,and the corresponding prediction model is proposed.Fourthly,contrast the measured data,use MATLAB to draw the model subsidence curve,and then make a comparative analysis.The main research results of this paper are as follows:(1)In the piecewise Knothe function model,the variation law of c,?,and _vw with the advance of working face is analyzed.The value of parameter c of the same surface point changes continuously with the advance of working face from the beginning of subsidence to the end of mining,and its change image is similar to normal distribution.the time point(?)at the maximum ground subsidence speed is related to the position of subsidence point,the advancing speed of working face and the initial subsidence time of monitoring point.the ratio of the subsidence value(_vw)at the maximum ground subsidence speed to the maximum sinking value at this point gradually increases with the working face moves forward under the condition of insufficient mining,and under the condition of sufficient mining,the ratio is 0.5.For these characteristics,based on the theoretical knowledge of the probability integral method,the solution method of three parameters is given.Then,the segmented Knothe function was optimized.(2)To deal with the problems of most of the existing dynamic prediction models only have single parameters,the ratio of the subsidence value at the maximum ground subsidence speed to the maximum sinking value at this point is fixed and the problem that the subsidence velocity and subsidence acceleration are not consistent with the actual surface subsidence velocity and acceleration,a dynamic prediction model is established based on the variogram theory.The specific method to establish the model is as follows:Based on the theoretical knowledge of probability integral method to deduce the time of surface subsidence which was put into the range of the variogram.In order to make the model conformed to the law of surface subsidence when described the subsidence,velcity and acceleration of the surface point changed with time,a model using different indices was combined with the subsidence data of the four mining areas to conduct a comparative analysis.It was found that when the model molecular index(b?c1)was 4,the expected sinking time was the closest to the time calculated by probability integral method,and the denominator index(b?c2)was proportional to the sinking time,so the best index of the model was:b?c1=4,b?c2 was the value to be fitted and then a dynamic prediction model based on variogram is proposed.(3)Taking Tunlan Coal Mine as an example,comparing the prediction accuracy of the optimized segmented Knothe function model,the prediction model based on the variogram and the six existing models,the comparison results show that the two dynamic prediction models for surface subsidence proposed in this paper have relatively high accuracy,and have good stability.