| Peat soil,clay and silt are widely distributed in Dianchi Lake Basin,which is a regional soft soil with poor engineering properties.It is of great engineering significance to understand the basic properties of soil and accurately determine the soil mechanical parameters.The physical properties of three kinds of soil were measured by field sampling,and the influence of organic matter content on the physical properties of peat soil was analyzed.Aiming at the special peat soil,the indoor one-dimensional compression consolidation and permeability tests are carried out,and the relationship among stable compression,consolidation pressure and overburden self weight of peat soil under various pressures is analyzed.The applicability of Terzaghi’s one-dimensional consolidation theory is tested and modified.In addition,a large number of test data of three kinds of soil mechanical parameters are collected,and the clustering analysis algorithm is introduced into the analysis of geotechnical parameters for the first time.In view of the traditional mathematical statistics method can not consider the influence of prior information,the Bayes method is used to derive the posterior distribution combined with prior distribution and sample distribution,and the geotechnical parameters are estimated posteriorly.Combined with Fisher information,the convergence value of geotechnical parameters is predicted when the sample size tends to infinity.From the perspective of large samples to test the rationality of parameter values under small samples.Taking a foundation pit excavation project in Kunming as an example,by changing the soil mechanical parameters,the change trend of deep horizontal displacement in the excavation process is simulated by finite element software,and the parameter results obtained by Bayes method are verified by combining with the actual monitoring data.The main research contents and achievements are as follows:1)The basic physical properties of three kinds of soil were measured.In addition,the influence of organic matter content on the relevant indicators of peat soil was analyzed.2)Through the one-dimensional compression consolidation-permeability test,the changes in the compression curve of peaty soils with different buried depths under different consolidation pressures are measured,and the difference between the consolidation pressure and the weight of the overlying soil and the final settlement are discussed.relationship.A method is proposed to calculate the dissipation of excess pore water pressure in Terzaghi’s one-dimensional consolidation theory,which considers the change of the void ratio and compressibility coefficient with the void ratio,and verifies the feasibility of the method.At the same time,the change law of permeability coefficient in the integrated stress field of consolidation and seepage pressure is analyzed.3)In the case of large samples,K-means clustering analysis method is used for the first time to analyze the three basic physical properties of peat soil:water content ω,natural weight y,and relative density ds.It is found that when the cluster number k=1,the sample frequency histogram can not well meet the common normal,lognormal or gamma distribution;when k≥2,there is always a class in the classification that can keep k=1 The main information of the sample distribution,and the number of this kind of samples is the largest,and the sample frequency distribution is closer to the conventional distribution.According to the standard deviation σ2 of normal distribution,the value range of geotechnical parameters is determined by using different interval length division principles,and the feasibility of the method is tested by comparing with the actual value.4)Bayes statistical method depends on posterior distribution,so the prior distribution in Bayes method and the calculation method of super parameter in posterior distribution are deduced.Bayes method can consider the prior distribution and combine with the sample distribution to deduce the posterior distribution.When the sampling is completed,the sample distribution is determined,and the posterior distribution largely depends on the prior distribution.Therefore,three different sets of prior distribution are deduced:the non information prior distribution,the prior distribution determined by prior information and the conjugate prior distribution.This paper selects the calculation method of the posterior distribution of the unknown parameters to be studied in the population distribution,analyzes the unknown parameters μ and σ to be studied,and combined with the engineering example,comprehensively compares and determines that the optimal posterior distribution is the posterior distribution in the case of conjugate prior distribution.5)In order to deduce the convergence value of the posterior mean value of soil mechanical parameters when the sample size tends to infinity,combined with Fisher information,the posterior distribution of the mean value μ of soil mechanical parameters can be approximated by a normal distribution independent of the prior distribution,and the convergence value of the parameters can be obtained when the sample size tends to infinity.The consistency and asymptotic normality of soil mechanical parameters are tested by the limit of the posterior distribution It is found that the convergence value of soil mechanics index falls in the Bayesian estimation interval,which fully proves the feasibility and reliability of the analysis method.6)It is found that when the sample size increases,the reliable convergence value of geotechnical parameters gradually approaches to the posterior mean value μ(x),which is related to the sample mean value x and sample size n to a certain extent.After obtaining the convergence value of each layer parameter,the minimum sample size required for accurate parameter value is found in turn,so as to achieve the most reliable value of geotechnical parameters at the lowest cost.7)A two-dimensional finite element analysis model is established for the subway foundation pit in Huxiang sedimentary soft soil site.The geotechnical mechanical parameters determined by Bayes method and classical statistical method are input respectively for simulation calculation.The comprehensive comparison and analysis of the deep horizontal displacement value of the foundation pit calculated and actually monitored shows that the displacement value calculated by Bayes method is more close to the actual deformation Finally,the rationality of Bayes method is verified. |
PDF Full Text Request