Study On Evolution Mechanism Of Rock Fracture And Rockburst Prediction In Deep High Stress Environment

Various underground engineering in China are gradually developing to the deep.Under the condition of deep high stress,the rock fracture is easy to evolve into nonlinear mechanical phenomena or disasters such as zonal disintegration,slabbing and slabbing rockburst,which seriously affects the construction safety and long-term stability of underground engineering.It is difficult to explain these deep rock mechanics problems by using traditional linear and continuous medium theory.From the perspective of discontinuous medium mechanics,studying the fracture evolution mechanism of engineering-scale rock under high stress conditions and further revealing the formation mechanism of different fracture phenomena can provide a theoretical basis for exploring the key influencing factors of dynamic disasters and carrying out disaster prediction.It is helpful to deepen the understanding of the mechanical behavior of deep rock and provide valuable reference for the safe and efficient construction of deep ground engineering.For this reason,this thesis is supported by the National Natural Science Foundation of China "Research on mechanism of structural-control type dynamic disaster initiation and monitoring&forecasting methods in deep underground projects"(No.51934003).By means of discrete element numerical calculation,laboratory test and machine learning modeling,the fracture evolution mechanism of engineering scale rock mass under high stress environment and rockburst prediction method are studied in order to provide theoretical basis for the study of mechanical behavior of deep rock and engineering practice.The main research work and results are as follows:(1)The two-dimensional discrete element method is used to construct a rock mass model that can reflect the meso-fracture effect.The intuitive characterization of the zonal disintegration phenomenon is realized.The generation mechanism and evolution regularity of the zonal disintegration phenomenon are studied.The zonal disintegration characteristics under different initial stress,lateral pressure coefficient and rock brittleness conditions are analyzed,and the influence of the number of joint groups on the fracture mode of surrounding rock is proved.By comparing the numerical results with the non-Euclidean geometric model under incompatible deformation,the variation regularity of the fracture zone width is revealed.(2)The effects of joint trace length,spacing,friction and bonding properties on the zonal disintegration of surrounding rock were systematically studied.By constructing the same mechanical parameter model,controlling the stress conditions and setting different joint conditions,the zonal disintegration and slabbing phenomena with the same fracture shape characteristics are compared and studied,and the relationships between two deep high stress fracture phenomena and influencing factors are revealed.A representation method of principal stress axis and principal stress magnitude is proposed to intuitively characterize the distribution of principal stress magnitude and direction after tunnel excavation.Based on this method,it is found that the existence of joints makes the principal stress deflection effect caused by excavation change locally.(3)The three-dimensional discrete element method is used to construct a threedimensional rock mass model.The process and mechanism of slabbing are explored from the perspective of fracture evolution under the three-dimensional original rock stress state.The characteristics and variation of surrounding rock slabbing under different original rock stress states are studied.The influence of model resolution,Poisson ’s ratio and post-peak behavior on slabbing is revealed.The DBSCAN algorithm is used to establish the relationship between macroscopic fracture and mesoscopic cracks,and then the quantification of slabbing behavior is realized by calculating the macroscopic fracture depth and crack angle.The representation method of principal stress axis and principal stress magnitude is extended to three dimensions,and the principal stress rotation of surrounding rock after tunnel excavation is visually displayed.The influence of principal stress change on fracture form,expansion direction and distribution range is found.(4)Unilateral limited uniaxial compression tests of slabbing brittle rock under different loading rates were carried out,and the macroscopic fracture and strength characteristics of slabbing brittle rock were investigated.Based on acoustic emission monitoring technology,the evolution process of instability and failure of slabbing rock specimen weas quantitatively characterized by analyzing the global amplitude-frequency characteristics and b value of AE signals.According to the test results,it is pointed out that the tangential stress accumulation rate during stress redistribution in the engineering site has an effect on the instability failure mode of slabbing surrounding rock.(5)Based on the mechanism of slabbing rockburst,the important influencing factors of slabbing rockburst are taken as the characteristic parameters,and the ensemble learning algorithm eXtreme Gradient Boosting is used as the basic algorithm to construct the slabbing rockburst prediction model.The hyperparameter optimization steps combined with grid search and cross-validation methods are proposed to improve model prediction performance.The model uses the softprob objective function to achieve different rockburst levels-probability prediction.The SHapley Additive exPlanations method was used to quantify the influence of each characteristic parameter on the occurrence of rockburst.Through engineering application and comparison with other machine learning models,the accuracy and reliability of the prediction method are verified.