土工强度参数变异特征及边坡可靠度分析中合理样本数目研究

发布时间：2018-01-12 18:19

本文关键词：土工强度参数变异特征及边坡可靠度分析中合理样本数目研究　出处：《西南交通大学》2017年硕士论文　论文类型：学位论文

【摘要】：边坡稳定分析中,以概率理论为基础、考虑各种不确定性影响的可靠度设计方法已进入实用阶段。该方法主要需要明确岩土设计参数的各种不确定性,根据不确定性来源可归纳为岩土体固有的参数变异性以及统计不确定性。传统上,在假设土体空间是均质的基础上,采用随机变量模型描述土体参数的变异性,但实际上由于土体的沉积过程以及应力历史不同,导致土体参数在空间上存在一定的变异性,单纯利用传统概率统计的方法忽略了土体空间不同点物理力学性质的差异,已无法满足对土体参数空间变异性做出客观分析与评价的需要;工程中通过勘察取样试验获取有限样本,利用小样本下的统计值对参数真值进行估计推断,但是选取多大的岩土样本抽样数量才能最大程度反映总体统计值仍没有具体的标准,样本数目的差异将影响统计不确定性的大小。基于铁路工程勘察资料,选取了5个以黏土为主的工程地质单元,分别采用概率统计方法和随机场理论分析了土体抗剪强度指标c、φ的点变异与空间变异特征,比较了两种理论分析结果的差异,讨论了考虑参数c、φ空间自相关的必要性,并基于灰色关联度研究了 c、φ空间变异性与其他物理力学指标的关联特征;以计算机随机抽样方法模拟实际工程勘探取样试验,分别构建较为完备的土工参数的样本统计值空间及对应的可靠度指标结果空间,对比分析不同样本容量对可靠指标离散性的影响,探讨了大、中、小3种参数变异水平下样本数目的合理取值问题。论文研究主要取得了如下成果:(1)以传统概率统计方法和随机场理论为研究手段,分别计算了黏性土 c、φ参数的点变异和空间变异系数。研究表明,两参数的点变异系数与空间变异系数差异性较大,相较于点变异系数,c、φ值的空间变异系数折减率分别可达25%和15%。随机场理论通过方差折减的方法对强度参数的点方差进行折减,得到土体的空间方差,从而将点变异系数转化为空间变异系数,更真实地反映了参数的变异特征。(2)基于灰色关联理论探讨了液性指数IL、标准压缩系数a1-2、含水率w、天然孔隙比e和塑性指数IP等主要物理力学参数对抗剪强度指标空间变异程度影响的排序规律。分析表明,液性指数IL与c、φ参数的空间变异性关联程度最大,其次为结构性指标a1-2和含水性指标w。(3)针对样本数目与土工参数变异性影响可靠度分析结果离散程度的问题,提出了以控制可靠度分析结果空间变异程度反推样本数目的方法,探讨了不同土工参数变异水平下对应的合理取样数目。结果表明,可靠指标结果空间变系数取0.30、0.25、0.20时,对应于参数呈小变异水平的样本数为6、9、15,中、大变异水平时样本数目宜分别提高至 11、15、20 和 18、20、30。
[Abstract]:The slope stability analysis, based on probability theory, the reliability design method considering the influence of uncertainties has entered the practical stage. The main method to clear the geotechnical design parameter uncertainty, according to the sources of uncertainty can be summarized as parameter variability and statistical uncertainty inherent in rock and soil. Traditionally, in the assumption that the soil space is homogeneous on the basis of the stochastic variable model to describe variability of soil parameters, but in fact due to the deposition process of soil and the stress history of different soil parameters, cause there is variability in space, ignoring the differences of soil spatial differences of physical and mechanical properties of the simple use of traditional methods of probability statistics. Has been unable to meet the need to make objective analysis and evaluation of parameters of soil spatial variability; engineering survey by sampling test samples with limited access. Statistical small sample values for the true value of the parameters for estimation, but the selection of samples of rock and soil in order to maximize the number of sampling statistics reflect the overall value is still no specific standard, different sample number will affect the statistical uncertainty. Based on survey data of railway engineering, selected 5 in clay engineering geological unit, using probability statistics method and random field theory of soil shear strength index C, feature point variability and spatial variability of phi, compared the analysis results of the two theories, discussed by considering the parameters of C, with the necessary spatial correlation, and based on the grey correlation degree of C association with characteristics of spatial variability and other physical and mechanical indexes; computer simulation method of random sampling sampling test of the practical exploration, the sample statistics of geotechnical parameters of complete construction of value space The results of the reliability index and the corresponding space, comparative analysis of the influence of different sample size on discrete reliability index, are discussed, in the number of samples to the reasonable value problem of small 3 parameter variation level. This thesis mainly achieved the following results: (1) based on the traditional statistical method and random field theory research means of cohesive soil were calculated for C, coefficient of variation and spatial variation with parameters. The results show that the two parameter difference coefficient of variation and spatial variation coefficient is larger, compared to the variation coefficients, C, spatial variation of the value number reduction method rate can reach 25% and 15%. respectively by random field theory variance reduction of strength parameters of variance reduction, spatial variance of soil, and the variation coefficients into spatial variation coefficient to reflect the variation of parameters. (2) based on Grey Relational Theory To investigate the liquidity index IL, standard compression coefficient A1-2, w water content, natural void ratio e and the main physical and mechanical parameters such as plasticity index of IP against order of shear strength index spatial variability effects. The analysis shows that the liquidity index of IL and C, the spatial variability of correlation parameters followed by the maximum diameter. The structural index A1-2 and water content index of W. (3) according to the number of samples and the effect of soil parameter variability results of reliability analysis of discrete degree of the problem, put forward to control the results of reliability analysis of spatial variability of the sample number of backstepping method, discussed the rational sampling number of soil parameter variation under different level of the corresponding results. The results show that the reliability index of space variable coefficient 0.30,0.25,0.20, corresponding to a number of small sample parameter variation level is 6,9,15, when the number of samples should be large variation levels were increased to 11,15,20 and 18,20,30.

【学位授予单位】：西南交通大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TU43

【参考文献】