| Along with the thorough development of rock-mass engineering design theory, its probability design and reliability are paid more and more attention by people. But the rock mechanics parameter needed in the present reliability calculation process is usually got from simple statistics of the experimental data through classical probability statistical theory, which is based on the frequency stability but restricted by the sample quantity. Meanwhile, in these statistical distribution forms, the value scope of random variable is always from -∞to +∞or from 0 to +∞, which does not accord with the actual value of the rock mechanics parameter, and the calculation result often does not conform to the practical situation. In view of above-mentioned problems, the Bayes method of rock mechanics parameter inference and reliability under cutting-off-tail distribution were studied systematically in this paper. The main research contents and results are as follows:Based on the two types of errors theory of the probability statistical method, the paper systematically discussed the acquisition and processing of the prior information and the determination of the prior distribution etc., and established the calculation formula of the prior information reliability in the Bayes method of rock mechanics parameter inference.Based on the two types of errors theory of the probability statistical method and as to the different prior information and its different distribution form, the paper analyzed the influence on the post distribution due to the selection of different prior distribution, and established the robustness calculation formula and test method in the Bayes method of rock mechanics parameter inference.Using 50 uniaxial compressive strength, 50 uniaxial tensile strength, 50 elastic modulus and 50 Poisson's ratio history data collected from KANGJIAWAN Lead, Zinc and Gold Mine, as well as 10 uniaxial compressive strength and 10 uniaxial tensile strength tentative data, the paper inferred the rock mechanics parameter of KANGJIAWAN Lead, Zinc and Gold Mine as the prior distribution is normal distribution, log-normal distribution, extreme value, etc., and the cohesive force and internal friction angle by combining the 50 uniaxial compressive strength and 50 uniaxial tensile strength on Bayes method simultaneously. Furthermore, based on the hypothesis test theory, the paper proved that the rock mechanics parameter of KANGJIAWAN Lead, Zinc and Gold Mine obeyed normal distribution.In view of the characteristic that the mechanics parameter in rock mass engineering could not be negative and infinite, according to the stress - strength interference model in the rock-mass engineering reliability calculation, the cutting-off-tail reliability calculation formula of surrounding rock in tunnel was inferred when the stress and intensity were under normal distribution, log-normal distribution, exponential distribution and Weibull distribution.The finite element model of stress calculation of surrounding rock in transportation tunnel at the fourth level -280 of KANGJIAWAN Lead, Zinc and Gold Mine was established on the basis of the rock mechanics parameter inferred by Bayes method. The reliability of surrounding rock in the tunnel was calculated by Monte-Carlo LHS method, and the result showed that the reliability indexes of surrounding rock in the tunnel were all greater than 3.7. The operating of the transportation tunnel proved that the stability of tunnel was good, which showed that the parameter inference was reasonable, the reliability analysis result was in line with practical situation of this mine.The research in this paper is based on the subject frontier. On the basis of the latest mathematics method and advanced experimental method, the Bayes method of rock mechanics parameter inference and reliability under cutting-off-tail distribution were studied in this paper. The research of this paper further enriched and developed the rock -mass engineering and reliability theory, and improved the reliability calculation method of rock-mass engineering, which provided a scientific basis for the reliability analysis and design of rock-mass engineering. The research had an extremely important theoretical significance and engineering practical value for reducing the experimental quantity, reasonably using construction resource and so on.
|
PDF Full Text Request