基于随机场理论的顶管隧道施工地表变形特性分析

Analysis of surface deformation characteristics of pipe-jacking construction based on random field theory

针对顶管施工引发的地表变形问题,采用基于高斯型自相关函数和Karhunen-Loeve(KL)级数展开法的随机场理论,考虑地层参数的空间变异性及自相关性,建立三维数值模型中土体参数的随机模拟方法。结合具体顶管施工案例,分析地表纵向及横向变形规律,并与现场实测结果进行对比验证计算方法的可靠性。研究地层土体摩擦角、黏聚力、弹性模量的变异性、相关距离和变异系数对地表变形的影响规律。研究结果表明:与确定性数值模拟相比,随机模拟方法在分析顶管施工地表变形时具有更好的适用性;三维随机场较二维随机场更能够反映3个维度的空间变异性和顶管顶进的空间效应,得到不同顶进距离下各断面地表变形超出控制值的概率;弹性模量的空间变异性对地表变形随机曲线的离散程度和包络范围有较大影响,但不改变变形趋势和规律;变异系数和相关距离对随机曲线的离散程度有较大影响。

For the surface deformation problem caused by pipe jacking construction,a random field theory based on Gaussian-type autocorrelation function and Karhunen-Loeve(KL)level expansion method was used to establish a random simulation method of soil parameters in the 3D numerical model considering the spatial variability and autocorrelation of ground parameters.Combined with specific pipe jacking construction cases,the longitudinal and transverse deformation law of the ground surface was analyzed,and the reliability of the calculation method was verified by comparing with the actual measurement results on site.The variability of friction angle,cohesion,elastic modulus of stratified soil,correlation distance and coefficient of variation on surface deformation were studied.The results show that compared with the deterministic numerical simulation,the random simulation method has better applicability in analyzing the surface deformation of pipe jacking construction.The three-dimensional random field can better reflect the spatial variability of the three dimensions and the spatial effect of pipe jacking than the two-dimensional,and obtain the probability of surface deformation exceeding the standard for each section at different jacking distances.The spatial variability of elastic modulus has greater influence on the dispersion degree and envelope range of the random curve of surface deformation,but does not change the deformation trend and law.The coefficient of variation and correlation distance have greater influence on the dispersion degree of random curve.