Risk Characteristic And Assessment In A Long Slope Mining Construction Excavated By TBM

China is one of the largest coal production countries, coal will be the dominant resource in quite long time in future. With the increasing excavation and construction technique and improvement of design theories in coal development of our country, it has great superiority using slope under the condition of shallow and buried depth. Using the slope for excavation can save cost and enhance the efficiency, meanwhile, the construction investment is fewer and speed is fast, so TBM will be more and more widely used in coal construction.Drilling and blast method cannot increase the construction speed of long slope in mining, the TBM technique is suitable for long slope mining construction. It has the advantages of highly mechanization and automation and can manifest its superiority in long slope mining construction. Therefore, the prospect of TBM technique in mining construction will be more and wider.In recent years, China continuously achieve major breakthrough in TBM design, and solve technical challenge. It fills the gaps of TBM construction and develops by leaps in these years. The diameter of TBM section has gradually increased so it can adapt many kinds of section. The construction technique has been improved and the geological adaptability is enhanced in order to meet the need of actual construction. It has a great significance to promote the development of coal resources.Long slope mining construction using TBM is an open system while is affected by the surroundings such as staff, material and nature factors. The engineering techniques are well developed and the non-engineering techniques are deficiency especially the techniques that can reduce risk of construction by TBM. So it needs to set up the corresponding risk control model and algorithm in order to extract nonlinear characteristics of time series. Analysis, prediction, establish and feedback control of risk.Therefore, the suitable methods is needed to estimate, predict and control to provide a scientific foundation combined with actual construction. There are many factors influenced the risk which generation mechanism is complicated. Risk analysis of TBM construction is still in its infancy, there is few experts in the area of risk analysis on long slope mining construction by TBM, there are many classical methods in risk assessment and prediction. These method are the basis of controlling and management in long mining construction excavated by TBM. However, the method of quantitative analysis are scattered, the research depth is very limited which did not form relatively perfect theory system. Other subjects has made great progress in the area of model establishment and research methods, these progress can be used as beneficial reference for controlling and management of risk in mining construction by TBM. Therefore, this dissertation introduces nonlinear dynamic analysis methods and tries to analyze the changing mechanism. Set pair analysis, chaos prediction theory, multi-fractal method and multi-scale entropy method are used to assessment, prediction and fluctuation analysis. New progress has been made as follows:(1)Risk analysis and identification method of long slope mining construction excavated by TBMThe geomorphology of Taige Temple mine affiliated to Shenhua Xinjie energy enterprise is typical desert—semi-desert of plateau. The surrounding rock of mining area is mainly coarse grained sandstone. The mine lies in mid-temperate zone, and there is severe sandstorm. There are four distinct seasons with a significant climate change. It belongs to arid—semi-arid continental monsoon climate. There is no surface water, and the seismic intensity is Ⅵ which means weak seismicity region.TBM is classified by the geological conditions of surrounding rock, shield form, the diameter of TBM, sectional form of excavation and horizontal-vertical degree of tunnel. According to the definition of risk, the characteristic of risk is analyzed which presents objectivity, uncertain, well-management and linkage. These characteristics determine that the risk is uncertainty, so controlling and management models are needed to analyze the risk.According to the definition of risk, the risk in long slope mining construction excavated by a TBM can be categorized into five levels. In this paper a 10-point scale method is used to further describe the levels of risk. The higher the score is, the greater effect of risk. The two-layer risk assessment system is established, the first layer includes six categories of risk: nature risk, geologic risk, technique risk, equipment risk, management risk and accident risk. The second layer is more detailed, including 23 risk factors.(2)The risk assessment of risk appetite in long slope mining construction excavated by TBM(1)The risk assessment in long slope mining construction excavated by TBMThe set pair analysis can analyze the uncertainty characteristic from the system point, depicting the dynamic changing process of risk. Seven connection degree of set pair analysis is proposed and triangular fuzzy number is introduced, the assessment of risk is established and the values of b1、b2、b3 are discussed and improved. The coefficients are generalized to widely meaning where set to x,so the expression of seven degree is obtained generally.Entropy weight method is used to determine the index weight of risk factors, and the assessment model of risk is established.(2)Assessment model of risk with appetite in long slope mining construction excavated by TBMThe concept and category are summarized. The risk appetite is the different attitudes towards risk assessment. The different attitudes are mainly about the different judge of severity of risk, including optimistic attitude, neutral attitude and cautious attitude. The staff with optimistic attitude hold the opinion that the risk is objective, the value is also objective when judgment; the staff with cautious attitude hold the opinion that the risk is easy to influence the construction, the value is higher when judgment.This dissertation analyzes the reason of risk appetite with features of TBM construction, the expression between risk value and connection degree of x is given, and the risk value range of staff with different risk appetites are obtained.The results shows that: the risk value range of staff with cautious attitude is [0,0.5),the risk is higher; the risk value range of staff with optimistic attitude is(0.5,1.0],the risk is lower; the risk value range of staff with neutral attitude is 0.5.(3)The trend analysis of in long slope mining construction excavated by TBMThe theory of partial connection number and set pair trend can analyze the trend of system changes from the perspective of the development. According to the definition of partial connection number and set pair trend, the identity-discrepancy- contrary(IDC) model is set up to discuss the research approach for risk development tendency of mining construction excavated by TBM. The research needs to focus on the discrepancy and contrary, the factors in contrary is improved gradually. The system can adjust to the lower risk. the partial connection degree of the first order, second order, third order and four order of 23 risk factors are obtained and the development tendency of risk factor is given to understand the risk change from macroscopic view.(4)Dynamic multivariable risk prediction of long slope mining excavated by TBMChaos prediction has been widely used in many trades; it is mainly about predict the nonlinear time series. Chaos system presents random, but through phase-space reconstruction the regularities of system can be extracted so as to depicting the nonlinear mechanism of inner system. Therefore, chaos focus on the changing process and can be used to analyze the nonlinear characteristic of risk time series.This dissertation reconstructs phase-space of risk time series of TBM, recovering the original chaos system from the data. The combination method of first order local prediction and double-layers BP neural network is used to predict the risk time series of TBM. Firstly, principal component analysis is used to simplify the risk factors, four components are classified from 23 risk factors, the total rate of contribution is 92.7255%. Also, the loading matrix and expression of principal component are obtained.The risk factors that influencing the risk are analyzed and the phase-space of four risk time series are reconstructed. Time delay and embedded dimension is calculated and the largest value of lyapunov is above 0, it displays that the chaos prediction can be used to predict it.The result fits the actual risk well, so this approach has a certain theory to risk predict the risk of using TBM for construction.(5)Multi-fractal analysis on risk fluctuation of long slope mining construction by TBM(1)The multi-fractal characteristics of risk fluctuation of long slope mining construction by TBMMulti-fra ctal is mainly about the self-similarity of system, that means there has self-similarity of partial and whole in system under a certain extent. Any section of mutual independence is the epitome and represent of system. The theory of multi-fractal focus on the dependence of part with the whole. The theory of multi-fractal can deeply analyze the relation of partial change with the whole change.Mann-Kendall method is used to determine risk time series fluctuation characteristics. The curves of serialization and de-serialization data series re obtained, there is obvious risk fluctuation according to the regularities of curves. On this basis, the box-counting method is used to estimate fractal dimension of different risk time series of different levels. The points of data can present a straight line which means there is a fractal characteristic of risk time series. The approach of Hurst index is used to identify multi-fractal characteristic of five kinds of time series. Hurst index curves have obvious fluctuation, moreover, the changes of h(q) are greater when q < 0 than those of h(q) when q >0 which proves that risk time series fluctuation have multi-fractal features, so the approach of multi-fractal can be used to analyze the risk time series.Multi-fractal spectrums of five risk time series are obtained and the results shows that the greater the risk abrupt changes, the bigger the fractal intensity; the slower the risk abrupt changes, the smaller the fractal intensity.Another two parameters are proposed to describe risk multi-fractal complexity: Q and l, so as to fully extract the multi-fractal feature of risk time series and analyze the complexity of risk fluctuation. The result shows the greater intense the risk fluctuation, the bigger the Q and l; the less intense the risk fluctuation, the smaller the Q and l.(2)Multi-scale entropy feature of the time series of an abrupt change of risk in a long slope mining construction excavated by TBMMulti-scale entropy algorithm measures complexity change of time series in different time scale. It can display the fluctuation mechanism of time series and contained the dynamic information. This dissertation use the multi-scale entropy algorithm to measure the fluctuation mechanism of risk, three kinds of different time series of risk abrupt change is analyzed:A:the risk level steadily increases or decreases after an abrupt change with the same risk level before the change, including:a1:The risk level steadily increases after an abrupt change with the same risk level before the changea2:The risk level steadily decreases after an abrupt change with the same risk level before the change, including:The multi-scale entropy algorithm can be used to better analyze this kind of the abrupt change of the time series. The risk level of an abrupt change has a negative relation with the complexity of the multi-scale entropy; multi-scale entropy cannot effectively distinguish whether the risk level of an abrupt change is higher or lower.B:Risk levels are different before and after an abrupt change, with the same level of differencesMulti-scale entropy algorithm cannot provide a good analysis of this case, for which the risk levels are different before and after the abrupt change, with the same level of differences. It displays that the multi-scale entropy algorithm can distinguish the relative extent of data series change not the value of data before and after change.C:Risk levels increase or decrease step by step after an abrupt change with the same risk level before the changec1:Risk levels increase step by step after an abrupt change with the same risk level before the changec2:Risk levels decrease step by step after an abrupt change with the same risk level before the changeMulti-scale entropy curve can better analyze an abrupt change of a risk time series, but it cannot distinguish whether the change is increasing or decreasing.The cumulant of differences method is used to thoroughly describe the nonlinear characteristics of the risk time series in a long slope mining construction. The cumulant of differences is defined, and the curves of cumulant of differences are analyzed under the case A、C. The cumulant of the multi-scale entropy differences has a negative relationship with the abrupt change of risk time series. The multi-scale entropy algorithm can quantitatively depict the change of the risk time series.