A Method for Predicting Subsidence in Thick Loose Layer Mines Based on Segmental Weighted Parameter Assignment
As a generalized subsidence prediction method for coal mining,the probability integral method has the prob-lems of too fast edge convergence and inaccurate subsidence prediction in the application of thick loose bed mining area,which brings certain difficulties to the"three downs"coal mining and environmental protection in thick loose bed mining area.To ad-dress this,Based on the probability integral method,this paper proposes a subsidence estimation method for thick loose layer mining area based on segmented weighted assignment,the monitoring points are divided into 2 sections,inner and outer,accord-ing to the location of the working face,obtaining the corresponding estimation parameters,establishing adaptive assignment function models to weight the combination of the estimation results.On this basis,in order to improve the accuracy of parameter inversion,parameter inversion research is carried out based on the standard particle swarm algorithm and the adaptive simulated annealing particle swarm algorithm,respectively.At the same time,in order to improve the performance of the assignment mod-els,two kinds of adaptive assignment models are established to carry out the model optimization.The test results show that:compared with the probability integral method,the subsidence prediction value calculated by the method for predicting subsid-ence in thick loose layer mines based on segmental weighted parameter assignment is closer to the measured data,and the re-sidual sum of squares of the two assignment models are 0.007 1 m2 and 0.005 8 m2,which are smaller than that of the proba-bility integral method(0.020 3 m2);compared with the two kinds of assignment function models,the model two is more accu-rate and more effective;under the same conditions,the accuracy of the adaptive simulated annealing particle swarm algorithm is higher than that of the standard particle swarm algorithm.This paper verifies the feasibility of the mining area zoning section prediction,confirms that the method can effectively solve the problem that the edge convergence is too fast in the application of probability integral method in the thick loose layer mining area,and has certain reference significance to the subsidence predic-tion of the thick loose layer mining area.
万方数据
维普
