Prediction Of Residual Deformation In Mining Subsidence Area Based On Combination Algorithm

Since the residual surface deformation in the mining place is particularly complicated and affected through a range of factors,the surface subsidence deformation will continue to occur for a long time after the surface becomes stable.The subsidence area is characterized by abruptness and concealment,so that the residual deformation of mining subsidence area will affect or even destroy buildings,and cause hidden risk to people’s property safety.Therefore,it is very vital to monitor and predict the improvement tendency of residual floor deformation in goaf.This thesis takes Ⅲ 1308(top)working face in Yuecheng Mining area of Shanxi Province as an example to predict and analyze the residual deformation in mining subsidence area.Based on the above background,the primary work content and achievements of this thesis are as follows:(1)Aiming at the hassle that the fundamental artificial electric field algorithm has insufficient exploration capacity and is convenient to fall into nearby most effective solution,The chaos strategy and the reverse learning strategy are introduced into AEFA,and the quality of the populace is improved,and the chance of the algorithm which jumps from the neighborhood choicest solution is expanded.The greedy strategy can make the population get the optimal value quickly.A simulation experiment using a benchmark function is carried out,IAEFA is used to clear up the traveling salesman problem.The evaluation effects exhibit that the elevated algorithm has higher reliability than different algorithms.Finally,IAEFA is used to clear up the traveling salesman problem.The evaluation effects exhibit that the elevated algorithm has higher reliability than different algorithms.(2)Based on the measured data of surface movement of Ⅲ 1308(top)working face in Yuecheng Coal Mine,Shanxi Province,different meta-heuristic optimization algorithms are combined with support vector regression model,and different combined support vector regression models and measured data are analyzed and predicted.In the analysis of the surface movement monitoring data of the Ⅲ 1308(top)working face,the accuracy of the IAEFA-SVR model is 90%,the accuracy of the PSO-SVR model is 70%,and the accuracy of the SVR model is 50%.Using the IAEFA-SVR model for prediction data,most of the maximum difference is 5mm to 20 mm,with high accuracy.(3)In view of the residual deformation after mining,this thesis uses the residual deformation prediction theory to predict and analyze the settlement after mining,and verifies the feasibility of IAEFA-SVR mannequin in predicting the residual deformation of mining subsidence area.The effects exhibit that the residual deformation estimated by way of IAEFA-SVR accords with the development law of the predicted theoretical values and meets the boundary requirements of the settling velocity in the decline period.By comparing the prediction impact with the theoretical outcomes of residual deformation prediction,it is found that the difference between the estimated cost and the theoretical value of 90% is inside 1mm.Compared with the PSO-SVR model and the SVR model,the prediction accuracy of each single measurement point is improved by at least 110%.The accuracy shows that the IAEFA-SVR model has good applicability in the prediction and analysis of residual deformation in this area.