Research On Deformation Mechanism And Settlement Prediction Under Thick Loose Layers Based On SBAS-InSAR Technology

Huainan mining area has always been one of the mining areas with high frequency and concentrated mining in East China.Deep mining has led to the destruction of the original stress balance in the rock body,and the movement of rock strata has led to frequent surface subsidence.Exploring the evolution characteristics of rock layers under thick and loose layers caused by mining and accurately predicting surface subsidence is of great practical significance for subsequent disaster prevention and reclamation planning in coal mining subsidence areas.This paper first selects 86Sentinel-1A images from June 2019 to March 2022 as the data source,uses SBAS InSAR technology to monitor the subsidence of the 2212（1）working face in Xieqiao Mining Area,and obtains the cumulative subsidence value of the working face after the accuracy verification of the monitoring data combined with the field Levelling data,and analyzes the deformation characteristics of the three monitoring lines.Afterwards,based on the data of the working face,numerical simulation and calculation of surface movement and deformation parameters were conducted in FLAC3D to study the characteristics of overlying strata changes and surface changes during mining.Finally,the optimal combination of influencing factors was selected to construct a surface subsidence prediction model.The main research content and conclusions are as follows:（1）The average annual settlement rate within the working face during the monitoring target period is basically between-149mm/yr and-100mm/yr;The trend of deformation along the strike line from west to east shows a small cumulative settlement on both sides and a large cumulative settlement in the middle;The closer the points of the two dip lines are to the roadway,the greater the cumulative settlement;The deformation of the overall monitoring point first enters an active period,and then gradually enters a declining period.The maximum cumulative deformation values on the three monitoring lines are-409.127mm,-387.220mm,and-542.795mm,respectively.（2）Build a numerical model on FLAC^3Dand simulate and analyze the changes in the internal stress field,displacement field,and plastic zone failure field of the overlying strata under different excavation distances.It is found that the stable stress arch formed by vertical stress continuously moves forward,the concentration range continues to increase,and the vertical stress also shows an increasing trend;Horizontal stress is transmitted to both sides of the overlying strata,forming stress concentration at both ends of the goaf.The closer the goaf is,the more obvious the phenomenon of horizontal stress concentration is;The vertical displacement increases with the increase of the advancing distance,and the closer the displacement is to the excavation position,the more significant the change is;The failure of the plastic zone is distributed in an inverted saddle shape,and during mining,the weight of the overlying rock layer in the goaf shifts towards the new support points around the goaf,causing the overlying rock layer to collapse.（3）The surface movement and deformation parameters of the target working face during the research period can be obtained as follows:the starting distance is 27.65m;The strike boundary angle is 61°;The boundary angles of the two dip lines are66°and 56.4°;The maximum sinking angle is 87.1°;The maximum sinking speed can reach 0.531mm/d,showing a trend of first decreasing,then increasing,and then gradually decreasing;The maximum subsidence velocity hysteresis distance on the surface is 67.5m;The maximum sinking velocity hysteresis angle is 84.9°.（4）After Pearson correlation analysis,the main causes of ground deformation were analyzed.Finally,six influencing factors such as propulsion distance and the settlement value of ML40 point were selected as training input parameters for the prediction model.After selecting the Stacked LSTM model and GWO optimization algorithm to optimize the basic LSTM model,construct the GWO Stacked LSTM surface subsidence prediction model.Figure[40]Table[14]reference[98]...