Prediction Of Damage Level For Buildings Adjacent To The Mining Area

As a country with a large coal production and consumption,frequent underground mining and open-pit mining activities have caused the movement and deformation of the ground surface.Due to the tight land space,the movement and deformation of the surface spread to the surrounding buildings.The surrounding areas of the mining area are mostly masonry and frame buildings.It is sensitive to the deformation of the ground surface and is prone to structural damage.In order to prevent the loss caused by the surface deformation and predict the damage level under the surface deformation of the building,compared with the performance index as the index to measure the damage level,this article starts from the general cause of the damage of the building and adopts the limited tensile strain method.As a standard,the deep beam model is used to study the maximum tensile strain caused by curvature deformation and horizontal deformation,and sensitivity analysis is carried out using building attributes and ground surface deformation values as influencing factors,taking into account the relative stiffness ratios and influencing factors of different types of buildings,The building is divided into masonry and frame structure,while considering the prediction source of the ground surface deformation value,it can be divided into two cases: the ground surface deformation value under the building or the free site surface deformation value prediction.The free site surface deformation needs to consider the mutual soil structure Function and mechanical properties of foundation soil.Based on the specification and actual empirical parameter value ranges,a theoretical database was established for the above four situations,and the SVM model was trained to achieve prediction accuracy,and the f1 value of the prediction model was greater than 0.85.For complex data acquisition,the SVM model is used for single factor missing analysis to select important factors and simplify the attribute characteristics.The model verification and analysis are carried out by using an actual building damage case in a mining area,and the damage level is predicted based on simplified attribute characteristics.Compared with the level determined by the real crack,the accuracy of the prediction level is 90%.The simplified SVM can quickly and effectively predict the deformation and damage level of the building's surface.