Study On Large Deformation Prediction,Classification And Support Technology Of Jiumian Expressway Soft Rock Tunnel

With the shift of China’s economic construction center to the central and western regions,the construction scale of highway transportation and other infrastructure is becoming larger and larger,highway tunnels have been widely used,and China’s transportation network has been further developed.Due to the lack of understanding of the large deformation of tunnel surrounding rock under the combined action of multiple factors and the particularity and complexity of the material properties of weak surrounding rock,it is difficult to effectively predict the occurrence of large deformation and take effective support and reinforcement measures to prevent and control.Relying on the tunnel group of Jiumian expressway,this paper studied the rock mechanical characteristics of the tunnel site through mechanical experiments,on-site monitoring,statistical analysis,numerical simulation,machine learning and other methods,made large deformation grade prediction of the tunnel risk section in line with the on-site law,and further constructed a reasonable support system for the surrounding rock with different large deformation grades.The main research work and conclusions are as follows:（1）The indoor mechanical test results of rocks in the tunnel site show that the uniaxial compressive strength of rocks in different water bearing States is different,the softening coefficient of phyllite is 0.77,the water saturated rock sample has obvious through cracks along the weak bedding surface,and the failure surface is smooth;The triaxial compressive strength and shear strength of rock are anisotropic.The weakening effect of bedding weak structural plane is the most significant when the inclination angle is 45°.Combined with the numerical test results of triaxial compression,the strength anisotropy curve shows a"U"shape,and the numerical simulation is close to the mechanical experiment,which verifies the correctness of the mechanical experiment parameters.（2）Using the methods of on-site monitoring,statistical analysis and numerical simulation,the deformation law of surrounding rock and the mechanical characteristics of supporting structure in the Jiumian high-speed tunnel site area were explored,the standard for the classification of large deformation grades of the tunnel group of Jiumian Expressway was established,and the influencing factors of large deformation of surrounding rock in the tunnel site area were studied.The results show that the deformation of the surrounding rock and the stress asymmetry of the supporting structure in the tunnel area are significant,the deformation of the surrounding rock is mainly based on matrix shear failure and joint shear failure,the increase of ground stress leads to the nonlinear increase of the surrounding rock displacement and the minimum principal stress,the volume of the plastic area is linearly increased,the ground stress is controlled by the joint weak surface when the stress is low,and the stress concentration is controlled by the shear strength of the matrix when it is high,and the stress concentration part is transferred from the arch waist to the arch foot;the peak part of the surrounding rock displacement has follower to the weak surface of the stratigraphy,and the steep inclined rock mass is more prone to large deformation,and the minimum principal stress is more likely.The volume of the plastic zone increases first with the increase of the rock formation inclination and then decreases;the softening of the rock strength of groundwater is finally reflected in the macroscopic deformation level of the rock mass,and the maximum horizontal displacement ratio of the surrounding rock under dry,natural and water-saturated conditions is 1:1.32:2.27,and the volume ratio of the plastic zone is 2.2:1.08:1.（3）Using the method of machine learning,based on the in-situ stress inversion and field large deformation monitoring measurement information of Jiumian expressway tunnel,a GS RBF SVM large deformation prediction model suitable for the tunnel site area was constructed.The influence ranking of surrounding rock large deformation evaluation indexes is U₂>U₅>U₁>U₃>U₈>U₇>U₆>U₄.The compressive strength of rock mass has the greatest influence on the prediction accuracy of surrounding rock large deformation.The model is applied to the prediction of large deformation in tunnel site area.Combined with the comparison of the prediction results of strength stress ratio method,it is found that the prediction results of large deformation model are safe and the error is small,which confirms the reliability.（4）The support system under different large deformation levels of surrounding rock was formed by statistical analysis method.The support system is verified and optimized by numerical calculation.The key support parameters are optimized by Latin hypercube sampling and mathematical analysis method,and the combination of key support parameters with safety,reliability and high economic benefit is obtained.When it is applied to practical engineering,the average value of surrounding rock deformation is reduced by about 70%,The damage of supporting structure has been significantly improved.