Stress-Energy Principle-Based And Data-Driven Models And Their Applications For Rock Burst Risk Prediction

With the trend of mining depth from shallow to deep,the dynamic disasters represented by rock burst limit the safe mining and utilization of coal resources,and it is of great significance to realize the accurate prediction of rock burst risk.This theis focuses on the theme of "Stress-Energy Principle-Based and Data-Driven Models and Their Applications for Rock Burst Risk Prediction".The theoretical analysis,experimental research,numerical simulation and engineering application are comprehensively used to rock burst risk prediction model by numerical modelling-based and microseismic data-driven model,physics-based and data-driven model for rock burst risk prediction.The main conclusions are as follows :By analyzing the stress environment of the whole process of rock burst evolution,the true triaxial top-coal-bottom combination experiment was designed and carried out,and the rock burst pressure energy principle of "one generalized stiffness condition and two dynamic load induced effects" was reproduced.Among them,the essence of stiffness initiation is the stiffness condition under static load and the generalized stiffness condition under dynamic and static load.The essence of cumulative deformation-type dynamic load-induced impact effect is that when the pre-peak accumulated permanent plastic deformation exceeds the critical strain value of impact initiation,the initiation is similar to the impact failure of postpeak catastrophic point under quasi-static load.The essence of the stress-incremental dynamic load-induced impact effect is that the input dynamic load energy can overcome the dissipated energy generated during the development of the sample from the pre-peak stress state to the post-peak impact starting stress state.The relationship between microseismic damage parameters and physical and mechanical parameters of coal rock in numerical simulation is established,and the method of weakening the strength parameters of coal rock based on microseismic damage parameters is proposed.According to the concept of damage and failure in damage mechanics,considering the probability of damage under the condition of coal and rock damage,a method of degrading fracture network based on microseismic damage parameters is proposed.Finally,Numerical Modelling-Based and Microseismic Data-Driven Model for Rock Burst Risk PredictionBased on the probability density function and Weibull distribution function,the theoretical model and data-driven model are associated.Six Physics-Based models of discrete point,linear,regional stress concentration and stress damage are proposed,and a data-driven model including seismic wave CT inversion reconstruction,microseismic damage reconstruction and microseismic attenuation cumulative reconstruction is established.The rapid coupling superposition of geological structure and mining information and prevention and control information is realized,and the approximate real-time inversion of mining stress concentration coefficient in the process of coal seam mining is obtained.A comprehensive Physics-Based and Data-Driven Model for Rock Burst Risk Prediction method is proposed in turn,and the real-time prediction of rock burst danger area and danger level is realized.Based on the risk assessment value index of rock burst,the risk area and risk level of rock burst in 7302 working face of Zhaolou Coal Mine and 401106 working face of Hujiahe Coal Mine are continuously predicted in two time periods.The distribution of microseismic events in the next time period shows that there is a good correspondence between the main occurrence of microseismic events and the medium and above risk areas predicted by the strong mine earthquake and the risk assessment value.The accuracy of the risk assessment value index in predicting the risk area and grade of rock burst and the reliability of continuous real-time prediction of rock burst risk are verified.There are 61 figures,12 tables and 99 references in this thesis...