Prediction Analysis Of Residual Deformation Time Function Combination Model In Goaf Site

When the construction of projects such as high-speed railway is carried out in in goaf site,the accurate prediction of the ground subsidence dynamics and the residual deformation is expected to be particularly important for the standard construction as well as the safe operation and maintenance of the project.In response to the problems of limited applicability and low prediction accuracy and stability of the individual time functions in the dynamic prediction of surface subsidence in complex goaf,Arctangent time function and Tanh time function are proposed.A hybrid algorithm based on the Fruit Fly Optimization Algorithm(FOA)and the Simulated Annealing Algorithm(SAA)is developed for obtaining the global optimal solution of the time function parameters.And the optimal non-negative variable weight combination model with the minimum error sum of squares as a criterion is introduced,and the calculation method of the residual deformation of the goaf site is given on the basis of the surface subsidence prediction by combining the probability integral method.The following main conclusions are drawn from this paper:(1)Arctangent and tanh time functions are optimized for dynamic prediction,respectively.The influence laws on the surface subsidence,subsidence velocity and subsidence acceleration curve patterns of the two time functions when the parameters are changed are analyzed,and the spatial and temporal variation characteristics of the two time functions are discussed.Among them,Arctangent time function satisfies the basic characteristics of the ideal type time function,and Tanh time function also basically conforms to the objective law of surface subsidence.(2)Transforming the time-function parameter solution problem into an iterative optimization-seeking problem for the minimum of the objective function,the hybrid algorithm established in this paper quickly locks the interval of the optimal solution by FOA global search and perturbs the current solution locally with the help of SAA,so as to accurately identify the optimal solution of the parameter at the minimum of the objective function.And the hybrid algorithm has the advantages of simple operation logic,stable convergence state and high solution accuracy,which is suitable for solving the problem of accurate acquisition of time function parameters.The applicability analysis of the combination prediction model shows that in the prediction of surface subsidence based on time function,the error of variable-weight combination prediction is smaller than that of equal-weight combination prediction,and the optimal non-negative variableweight combination model is more superior in the variable-weight combination prediction,thus becoming the preferred combination model for dynamic prediction.And fifteen double term time function combination models such as Knothe-Weibull,Knothe-Logistic and Knothe-MMF are constructed.(3)Based on the basic principle of probability integral method,the probability integral model of horizontal or slowly inclined coal seam and rectangular working face is established,and the method of dividing the working unit of regular working face during dynamic mining is discussed.And the time function combination model constructed in this paper is combined with the corresponding values assigned to each working unit,so as to realize the dynamic prediction of surface movement deformation.On this basis,the expected model of residual deformation at any point of the surface of goaf is given,and the impact on the high-speed railway when the residual deformation of goaf site is too large is analyzed.(4)The dynamic prediction of working face 35101 shows that the optimal nonnegative variable-weight combination model extends the valid interval of the prediction curve by two time function curves near the measured settlement value,and the information of the single time function in the valid interval is filtered and reorganized with the help of variable-weight coefficients to obtain the optimal subsidence prediction curve that matches the measured subsidence value throughout.Among them,the combination model composed of Knothe function has better prediction performance among fifteen combination models,thus the combination model with the most suitable time function for the mining working face can be selected for the prediction of surface movement deformation and the prediction of residual deformation during dynamic prediction.(5)From the predicted results of 1176 east working face,it can be seen that the prediction results of surface movement deformation at any point at different moments are consistent with the theoretical development trend of surface movement basin.And from the residual deformation prediction results,it can be seen that the characteristics of residual deformation change along the working face strike direction are consistent with the objective law of surface movement deformation.And the residual subsidence change law at any point of the ground surface at different moments corresponds to the actual mining process,thus proving the feasibility of the residual deformation prediction model,and the surface stability can be reasonably evaluated by the residual deformation values at different moments.The residual deformation,as an important part of the foundation deformation of the mining area site,can provide a kind of basis for the feasibility evaluation of high-speed railway construction in goaf site.There are 50 figures,22 tables,and 113 references.