基于地层与土体参数变异性的深厚非均匀填土场地沉降特性研究

Due to the special topography and engineering properties of the loess area, excessive settlement of the high fill loess foundation has resulted in a large number of failures of the project. Substantive uncertainty of fill foundation is determined by the superposition of two kinds of soil with different geological causes and stress histories. Combining randomness of soil parameters with spatial location, and considering the variation of strata between undisturbed soil and filling layer, it is very important to study the characteristics of foundation settlement. Based on field and indoor test data of filling site, soil parameter random field model and geo-statistics theory are applied to study the variation law of parameters with soil depth, which aims to carry out the evaluation of spatial variability. Based on the coupled Markov chain model, random distribution and transfer state of the original soil layer and the filling layer in the region are obtained, and the formation variability is simulated. Considering spatial variability of ground and soil parameters, distribution of the spatial variation soil is mapped to the fill foundation model. Based on synergetic sequential gauss method discrete multivariate geotechnical parameter random field, combined with stochastic finite element method, foundation settlement at different depth is calculated, and random influence factors of fill foundation are analyzed, and variation law of reliability index of foundation settlement of deep fill is revealed, and the reliability analysis method of settlement prediction is put forward. The research results can provide a theoretical basis for the calculation and revision of the foundation design of the high filled loess site.

由于黄土地区特殊的地形地貌和工程性质，高填黄土地基的过大沉降已造成大量工程失效或失败。两种不同地质成因和应力历史的土体叠加在一起，就决定了填土地基本质上的不确定性，将土体参数随机性与空间位置结合，考虑原状土层与填土层间的地层变异性，开展地基沉降特性研究具有重要意义。本项目拟基于填土场地现场和室内试验数据，采用土体参数随机场模型和地质统计学理论，研究参数随土层深度变化规律，进行空间变异性评价。基于耦合Markov链模型，获得区域内原状土层与填土层随机分布转移状态，实现对地层变异性模拟。考虑地层和土体参数空间变异性，将空间变异土体分布映射至填土地基模型，基于协同序贯高斯法离散多元岩土参数随机场，结合随机有限元，计算地基不同深度处沉降，分析填土地基的随机影响因素，揭示深厚填土地基沉降可靠指标变化规律，提出沉降预测的可靠度分析方法。研究成果可为高填黄土场地地基基础设计计算、规范修订提供理论依据。

由于特定的生成环境和赋存环境，黄土地基在漫长的地质年代中经历了长期、多循回地质作用，填土地基由于丧失了原有土体的地质历史信息，填筑方式各异，决定了土体参数具有显著的空间变异性。本项目选取甘肃省兰州新区典型大厚度填土场地，开展了标准贯入试验、剪切波波速等现场试验和物理力学性质试验。分别针对兰州新区大厚度填土场地、西安、兰州、渭南三地黄天然黄土地层、宝中铁路高填路堤，基于随机场理论地质统计学理论，开展了土体参数去趋势化处理、平稳性与各态历经性检验、半变异函数的计算与拟合，建立天然黄土层和填土层的物理力学性质指标随机场模型。针对随机场模型中相关距离和变异性，采用递推空间法、半变异函数法等求解土体参数竖直方向的相关距离，探讨分析各土体参数间相关距离和变异系数存在差异的原因与主要影响因素。采用马尔科夫链蒙特卡洛模拟(MCMCS)生成收敛至后验概率密度函数的大量等效样本，以等效样本作为土体参数特征真值的估计统计样本，建立了土体参数概率模型。考虑填土地层变异性和土体参数的空间变异性，分别建立填土路堤有限元模型、填土地基条形基础有限元模型，采用协方差矩阵分解法得到土体参数随机场模型的离散参数值，将随机场模型与有限元模型网格的一对一单元映射，将随机场模型转换成数值模拟分析模型，提取每次随机有限元模型沉降变形，进行考虑土体参数空间变异性的沉降变形分析。针对天然黄土地层湿陷性在空间的变异性，结合Bayes理论和MCMCS生成大量等效样本作为土体参数统计特征的估计样本，计算细分间距至0.1 m的克里金自重湿陷系数预测值。将概率统计和随机场理论结合应用于湿陷量计算，结合模量折减计算方法和有限元数值分析，开展了基于土体参数空间变异性的黄土湿陷性评价。依托兰州地铁一号线单线隧道工程，开展了土体参数的空间变异性对隧道开挖后地表沉降和地表水平位移的影响分析，在此基础上讨论对盾构隧道地层沉降可靠度的影响。