Study On Rupture Evolution Law And Rupture Trace Prediction Model Of Overlying Soil Under Reverse Fault Dislocation

Since the formation of the earth,it has experienced billions of years of geological changes,leaving many faults in the crust.An active fault is one that is active today and is likely to slip again in the near future.However,for a long time,the research objectives of major scholars on active faults are mainly focused on the earthquake caused by fault activity and the influence of displacement on tunnels,roadbed and pipelines,while the research on the destructive law of overlying soil by reverse faults is relatively rare.Based on this,combined with laboratory model device test,finite element numerical simulation software ABAQUS and artificial intelligence algorithm-generalized regression neural network(GRNN),this paper carried out research on the failure mechanism of fault activity on overlaying soil and the evolution law of fracture track.1.A set of reverse fault simulation device was designed to study the development law of the soil failure track in the single-layer homogeneous soil site and the layered homogeneous soil site under the action of reverse fault,and the pressure of the overlying soil under the movement of reverse fault and the change of the displacement of the top surface were analyzed.The results show that due to the existence of breccia soil in the lower layer,the stability of soil will become worse,causing more serious damage to the top surface,and the upper cohesive soil layer inclined to the fracture like a plate.2.The numerical models of deformation and failure of single-layer homogeneous soil and layered homogeneous soil under reverse fault activity are established,and the applicability of the numerical simulation is verified by laboratory test results.The effects of different fault dip angles,different dislocation quantities and soil mechanical parameters on soil failure were studied.The results show that the model with smaller fault dip Angle and larger vertical dislocation quantity is more likely to reach the transfixion of plastic failure zone.The effects of density and Young’s modulus on the plastic failure of soil mass are relatively small,the effect of cohesion on the plastic failure of soil mass is basically negligible,and the effect of friction Angle on the plastic failure of soil mass is the largest,which generally shows a trend that the larger the friction Angle is,the less easy the failure is.3.In order to more clearly reflect the relationship between the development of soil failure and various influencing factors,a cross-validated generalized regression neural network(GRNN)was introduced to establish a prediction relationship between various influencing factors and the length of the plastic failure zone of soil.A prediction model based on the generalized regression neural network was proposed.