基于有限钻孔的土体成层结构及空间变异性研究

发布时间：2019-03-04 19:06

【摘要】：土体是在复杂地质构造运动和应力历史下形成的沉积物,这使得其在空间中不同位置的物理力学性质具有不同程度的差异,其在空间上通常具有不规则成层分布的特性,并且同一土层内部的岩土参数具有空间变异性。因此,岩土工程结构设计之前需要在建筑场地内进行地质勘探以得到所需要的地质资料。然而,通过地质勘探所获得的地质数据是离散的,不能完全确定场地内未勘探区域的地质数据。为了给工程设计提供充足可靠的依据,基于有限钻孔对土体成层结构及空间变异性进行研究很有必要。土体成层结构研究方面,空间插值法不能用于复杂的土层模拟,变差函数法虽然可以模拟复杂地层的分布,然而其所模拟的地层连续性较差。相比之下,耦合马尔可夫链模型则可以克服以上问题,但是该模型无法模拟具有不同倾斜方向的地层交界面,也不能准确模拟地层中的棱镜体。本文对耦合马尔可夫链模型进行了改进,编制了相关程序,并验证了程序的正确性。改进的耦合马尔可夫链模型能够较好地模拟地质剖面内不同倾斜方向的地层交界面及变异地层(如棱镜体)。基于耦合马尔可夫链模型研究了钻孔布置方案对地层模拟结果的影响,钻孔的位置对耦合马尔可夫链模拟复杂土体成层结构可能产生较大的影响,并且同一土层内部出现变异的概率会随着钻孔间距的增加而增加。本文还分析了获得几种常见的土体成层结构的稳定模拟结果所需要的钻孔数量。同一土层内的空间变异性可以通过随机场理论进行研究,然而随机场会高估土体的空间变异性,这是因为它对地质勘探点的数据仍然采用随机模拟的方式生成。因此,本文研究了能够充分利用场地内实测值,并有效减小岩土参数空间变异性的三维条件随机场方法,编制了相应程序并验证了程序的正确性。基于三维条件随机场的方法,本文还研究了钻孔位置和钻孔数量对岩土参数空间变异性的影响,发现当岩土参数的相关性符合指数平方型相关函数时,条件随机场内勘探数据的主要约束距离约为该方向波动范围的一半,并且条件随机场预测值的精度受预测点周围的钻孔数量以及岩土参数变异系数的共同影响。
[Abstract]:Soil mass is a sediment formed under the complex geological tectonic movement and stress history, which makes the physical and mechanical properties of different positions in space different, and it usually has the characteristics of irregular stratified distribution in space. And the geotechnical parameters in the same soil layer have spatial variability. Therefore, geological exploration should be carried out in the construction site before the structural design of geotechnical engineering to obtain the required geological data. However, the geological data obtained from geological exploration are discrete, and the geological data of the unexplored area in the site cannot be determined completely. In order to provide sufficient and reliable basis for engineering design, it is necessary to study layered structure and spatial variability of soil based on finite drilling. In the study of layered structure of soil, the spatial interpolation method can not be used to simulate the complex soil layer. Although the variation function method can simulate the distribution of complex strata, the continuity of the simulated strata is poor. In contrast, the coupled Markov chain model can overcome the above problems, but the model can not simulate the stratigraphic interface with different inclined directions, nor can it accurately simulate the prism in the stratum. In this paper, the coupling Markov chain model is improved, and the related program is compiled, and the correctness of the program is verified. The improved coupled Markov chain model can well simulate the stratigraphic interface and variation strata (such as prism) in the geological section with different tilting directions. Based on the coupling Markov chain model, the influence of borehole layout scheme on the simulation results is studied. The location of borehole may have a great influence on the layered structure of complex soil mass simulated by coupled Markov chain model. And the probability of variation in the same soil layer increases with the increase of borehole spacing. The number of holes needed to obtain the stability simulation results of several common layered soil structures is also analyzed in this paper. The spatial variability in the same soil layer can be studied by random field theory, but the random field overestimates the spatial variability of soil because it still generates the data of geological exploration points by random simulation. Therefore, this paper studies a three-dimensional conditional random field method which can make full use of the measured data in the site and effectively reduce the spatial variability of geotechnical parameters. The corresponding program is compiled and the correctness of the program is verified. Based on the method of three-dimensional conditional random field, this paper also studies the influence of borehole location and number on spatial variability of geotechnical parameters. It is found that when the correlation of geotechnical parameters conforms to exponential square type correlation function, The main constraint distance of exploration data in conditional random field is about half of the range of fluctuation in this direction, and the accuracy of conditional random field prediction is affected by the number of holes around the predicted point and the coefficient of variation of geotechnical parameters.
【学位授予单位】：哈尔滨工业大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TU43

【参考文献】