| Since the last century,coal,as the main energy source,has been responsible for national energy security and sustained economic development.Large-scale and high-intensity coal resource mining will cause a series of environmental hazards of mine geology,such as destruction of aquifer,landslides,surface subsidence,and destruction of structures.As a new type of active surface deformation monitoring technology,D-InSAR(differential interferometry synthetic aperture radar)has the advantages of all-weather,all-time,low cost,large coverage area,and high temporal and spatial resolution.In recent years,D-InSAR has been widely used in mine deformation monitoring by a large number of scholars.To make up for the shortcomings of traditional monitoring methods,D-InSAR technology provides a brand new method for deformation monitoring and prediction in mining areas.First,by analyzing the many shortcomings of traditional methods in the inversion of probability integral parameters,and drawing on the experience of intelligent optimization algorithms widely used in solving highly nonlinear functions,the firework algorithm improvement and its application in the inversion of probability integral parameters are carried out.Then,in view of the difficulties in the application of D-InSAR technology in three-dimensional deformation monitoring of coal mine,a study on the dynamic monitoring method for solving three-dimensional deformation of coal mine taking into account the characteristics of D-InSAR monitoring was carried out;in addition,aiming at the problem that the deformation observation results of conventional surface observation stations cannot estimate the boundary angles and movement angles parameters of loose layers and rock formations,a method for separating boundary angle parameters of surface movement basin based on robust estimation theory is proposed.Finally,in view of the problems in the accurate detection of the time and space characteristics of the goaf that restricts the development and utilization of the resources of the closed and abandoned mines,the research on the detection method for the time and space characteristics of the abandoned coal mine based on InSAR technology is carried out.Based on the above three aspects of research,the following research results have been mainly achieved:(1)An inversion model for probability integral parameters based on the improved firework algorithm is constructed(MIFWA).Based on the analysis of strengths and weaknesses of FWA,the convergence speed,the exploitation capabilities at the optimum,exploration capabilities on the global and false convergence eliminating were improved,and on the basis,IFWA(improved fireworks algorithm)was proposed.Furthermore,the mining subsidence and the probability integral method were integrated,and the IFWA-based inversion model for probability integral parameters(MIFWA)was constructed.Simulation experiments show that the MIFWA is better than the MFWA in accuracy and stability,and the averaged relative error of the estimated parameters by MIFWA is 0.15%approximately.Furthermore,the observation of deformation based on D-InSAR technology can also better integrate MIFWA for solving parameters,and MIFWA has better performance in resisting observation errors,the anti-error of the geological mining generalization and the missing points.Finally,MIFWA was used in solving the probability integral parameters in the 1414(1)working face of Guqiao Mine in Huainan mining area,China.The obtained probability integral parameters are as follows:q=0.97,tan ?=1.98,b=0.41,?=89.08,S1=-5.94,S2=-14.89,S3=51.66,S4=30.55,and the RMSE of fitting subsidence and horizontal movement is 107.13 mm.(2)A dynamic monitoring method of surface three-dimensional deformation of coal mine based on fusion of single sight D-InSAR and BK model is proposed.Taking into account the characteristics of D-InSAR that it is difficult to obtain large gradient deformations,the Boltzmann function model with good boundary fit was selected,and a BK model was constructed by the fusion of the Knothe time function;Then,due to the decoherence of the time-space baseline and the monitoring principle of D-InSAR,it can often only obtain the short-term,one-dimensional deformation of the surface along the LOS direction.Therefore,based on the projection relationship between the three-dimensional deformation and the deformation along the LOS direction,combined with the geological mining conditions of the target area,the fitness function with non-cumulative deformation was constructed;Finally,a dynamic monitoring method of surface three-dimensional deformation of coal mine based on fusion of single sight D-InSAR and BK model was proposed by introducing the firework algorithm.Simulation experiments show that the construction method can reliably and accurately inverted all dynamic prediction parameters of mining subsidence,and the relative error of the solved parameters was between 0.11%and 7.51%.The subsidence and horizontal movement monitored by the construction method were consistent with the real values.The results of the anti-error experiment show that under the influence of observation error or model parameter error,the overall effect of the solved parameters by the construction method was better,and the ability to resist is shown to a certain extent.The construction method was applied to the 13121 working face of Gubei Mine in Huainan,and The conversion of short-term D-InSAR observation to 3D deformation dynamic monitoring based on the construction method is realized.The reliability and scientificity of the dynamic prediction method for solving the surface three-dimensional deformation of coal seam based on D-InSAR technology has been verified.(3)A boundary angular parameter estimation of the surface moving basin based on the weighted iterative least squares is proposed.The observation equation is derived by taking the rock movement angle,the rock boundary angle,the loose layer movement angle and the loose layer boundary angle as parameters.Considering the possible gross error in the comprehensive moving angle and the comprehensive boundary angle,The method of Hampel and IGG alternative iterations is introduced,and the optimal value of the parameters is obtained iteratively,and the accuracy of the parameter is evaluated.The experimental results of Guqiao and Gubei Mine show that:the parameter estimation accuracy of Hampel method is better than 4° and the estimation accuracy of IGG method is better than 1°;the parameter estimation result based on IGG weighted iterative least squares method:?0=48.02°,?=64.47°,?0=48.19°,?=69.18°,?0=48.20°,?=70,01,?0=35.32°,?=49.92°?(4)An inversion method for spatiotemporal characteristics of abandoned coal mine based on earth observation technology is proposed.the historical mining activities of abandoned coal mine are characterized by underground eight spatiotemporal characteristic parameters,and then based on the dynamic probability integral model,the relationship between underground dynamic mining and ground surface response is ed into a mathematical model.Then,according to Boltzmann function fitting and the analysis of change trend and sensitivity of dynamic probability integral parameters,unknown dynamic probability integral parameters are converted into known quantities,and a space-time relationship model between underground spatiotemporal mining characteristics and ground movement and deformation is constructed.Finally,Taking into account the highly nonlinear characteristics of the model,an improved fireworks algorithm(IFWA)is introduced,and then the inversion method is proposed.Real data experiments show that the construction method has high accuracy in obtaining underground spatiotemporal characteristics,and provides a new idea for abandoned coal mines that need to invert historical mining activities(mining time,mining location,coal mining space information,etc.).The simulation experiment results found that the construction method can better integrate the D-InSAR deformation observation technology under the condition that the model parameters remain error-free.The inverted underground spatiotemporal characteristics have high accuracy,and the inversion sequence of underground mining activities is consistent with the actual(the relative error is in the range of 0?21.0%,with an average of 5.0%).Figure[40]Table[19]References[123]... |
PDF Full Text Request