Study On Failure Characteristics And Intelligent Prediction Of Rock Burst In Deep Underground Engineering

With the increase of excavation depth of underground engineering,the ground stress environment and physical and mechanical properties of rock mass become extremely complicated,especially the rock burst disaster induced by excavation.Rockburst is a dynamic phenomenon in which the elastic deformation energy accumulated in the rock mass is suddenly released after the rock mass is disturbed by excavation,resulting in the burst and ejection of surrounding rock.Rockburst,as a typical engineering geological disaster under deep high stress environment,causes serious safety hazards to engineering construction personnel and equipment,and the prevention and control of rockburst has become the first problem to be solved in the construction process.The determination of rock burst failure characteristics and scope is the scientific basis of rock burst prevention and the determination of rock burst prediction method and prediction model is the theoretical basis of rock burst prevention.Aiming at the problem of rock burst disaster,the main research content of this paper is as follows:1)The mechanism of rockburst generation is explained from the perspective of energy.Based on E.Hoek criterion and energy criterion of rockburst occurrence,FLAC^3D numerical simulation software is used to explore the influence of in-situ stress environment on the failure characteristics and intensity level of rockburst.2)Taking the Mine-by underground test chamber in Canada as the research object,a variety of constitutive models are used to simulate the failure characteristics of the chamber,and the simulation results of the failure characteristics of the chamber are compared and analyzed with the actual failure characteristics.m-0 rule based on Hoek-Brown empirical criterion is selected as the constitutive model to simulate the failure characteristics of deep hard brittle rock mass.Finite element method,boundary element method and Lagrange finite difference method are used to simulate the failure of experimental chamber,and verify the scientism,applicability and validity of m-0 rule.3)From the point of view of physical and mechanical properties of rock,energy storage characteristics of rock and structural shape of chamber,four discriminant indexes affecting rockburst were selected,including uniaxial compressive strength,tensile strength,maximum tangential stress of chamber and elastic deformation energy index,and relevant discriminant index data of rockburst were collected and sorted out through literature to build a sample library.Based on the detailed analysis of the correlation between the four discriminant index data,a genetic neural network model（GA-BP model）was established considering various influencing factors of rockburst.The applicability and effectiveness of the genetic neural network model to predict hard rock burst were verified through the model learning training and prediction.4)On the basis of m-0 rule,the destruction characteristics of surrounding rock in the circular chamber of Jizjou and Xinchang Preselection site in the underground laboratory of high radioactive waste are simulated,and the genetic neural network model is used to predict the risk of rockburst and the corresponding intensity level of the preselection site.By means of distributed monitoring,the numerical solution of the maximum tangential stress around the chamber is obtained.The numerical solution of the maximum tangential stress around the chamberσ_{θm ax},uniaxial compressive strengthσ_c and tensile strengthσ_t are taken as the discriminant indexes for rockburst prediction.The predicted results are compared with those of other rockburst criteria and prediction models.The risk of rockburst and corresponding intensity level of the pre-selected site are analyzed comprehensively.