Coalburst And Rockburst Classification And Control Technology Evaluation Based On Uncertainty Theory

With the exploration of mining at great depth,the frequency of geological hazards induced by artificial excavation is more frequent.Prediction and control of coal bursts and rock bursts,influenced by a large variety of uncertainty information,is considered as a complex engineering problem.Traditional evaluation methods are hard to remove the uncertainty information existing in the whole evaluation process,making the results are inconsistent with the actual situations.In order to reduce the disturbance of uncertain information to the evaluation process and further improve the classification accuracy of proposed methods,this study mainly focuses on the disadvantages of existing researches and systematically carries out researches from the following aspects.(1)In the context of the application and selection of current uncertainty algorithms are unclear,a systematical model review is carried out to analyze the development trend and entire application of each model.Then,the publication characteristics of uncertainty theory in mining engineering are encountered,aiming to analyze the degree of promotion of these models in this area.The results show that the current researches associated with uncertainty algorithms focus on its application rather than optimization.Meanwhile,some suggestions associated with membership functions construction are proposed to further optimize these models.(2)A group of 147 data were collected to enrich the available dataset and to further improve the evaluation reliability of coal burst liability based on set pair and multidimensional cloud model.An empirical classification method used for coal burst liability evaluation was introduced integrating the subjective knowledge of decision-makers.Meanwhile,Kriging method was applied to produce 2D coal burst liability classification chart based on Wet and Ke in view of its excellent performance on spatial interpolation.The models/methods established in this research shows great potential in coal burst liability classification in comparison with other empirical criteria.(3)In the context of the index system for rockburst evaluation is complex and the determination of membership is unreasonable,information entropy is used to extract valuable information and calculate the index weights from the available dataset;Seven linear/non-linear membership functions are introduced to remove the impact of uncertainty information to the rockburst evaluation integrating unascertained measurement theory.Simultaneously,two groups of evaluation models are developed to evaluate the spalling and rockburst integrating credible identification principle and set pair analysis.Additionally,empirical knowledge of decision-makers is also considered to develop an empirical rating model,which is validated by available dataset.The aforementioned models can reduce the impact of uncertainty information in evaluation process considerably,and further validate its reliability through the engineering cases.(4)In order to evaluate the reasonability of control technologies designed for rock bursts and coal bursts,and to further reduce the uncertainty of onsite operations,in this study,the combination model integrating set pair analysis and multidimensional normal cloud model,and unascertained measurement theory are used to evaluate the water infusion in coal mass and destressability via constructing reasonable index system and collecting dataset.The evaluation results indicate that the two uncertainty-based models not only can accurately classify the rock bursts and coal bursts,but also can evaluate the reasonability of control methods.Simultaneously,GUIs referring to geological hazards classification and control technology evaluation are developed for each sub-module,which also integrated into a comprehensive evaluation system,aiming to high-efficiently evaluate the engineering issues in underground.