基于非参数回归的土层参数反分析算法及其应用研究

发布时间：2018-03-30 19:06

本文选题：土层参数　切入点：反分析　出处：《昆明理工大学》2015年硕士论文

【摘要】：土层参数是岩土工程设计过程中的重要指标,一般通过室内试验获得,由于土层的自然变异性、试验条件、试验方法等方面的原因,试验结果往往具有离散性,与真实值有一定差异。再者,如果工程施工方法不同,即使是同一种土体,其土层参数也会存在差异。因此,需要采用适当的方法合理的确定土层参数。结合已有监测数据,通过反分析手段获得土层参数,然后用反分析得到的参数进行正分析,从而可以及时更新数据、指导后期设计施工或指导类似工程设计。本文从反分析方法及其应用方面展开研究工作,如下：(1)针对目前反分析方法存在的问题,提出了一种基于非参数回归的土层参数反分析算法。将不同工况的同一实测目标作为约束条件,考虑实测目标量测的误差,选定容差ε,从而得到目标实测值的容差范围,然后确定待反分析土层参数的初始范围,结合均匀试验和非参数回归方法建立土层参数与实测目标之间的响应面,按照工况枚举处大量参数样本,利用响应面插值得到大量目标计算值,逐一对比目标的计算值和实测值,筛分出在容差范围内与目标实测值接近的目标计算值所对应的土层参数数据样本,并把数据样本的统计范围作为目标实测值在一定容差范围内对应的土层参数实际取值区间。(2)反分析算法在基坑工程中的应用。选择钢支撑力和桩身水平位移作为实测目标,利用反分析算法对基坑剖面支护桩所在深度范围内土层的抗剪强度指标c、φ值进行反分析,得到土层参数的取值区间,并进行了验证。(3)反分析算法在盾构隧道工程中的应用。选择地面沉降作为实测目标,利用反分析算法对盾构隧道开挖影响范围内土层的抗剪强度指标c、φ值进行反分析,得到土层参数的取值区间,并利用反分析结果预测临近将要开挖断面的地面沉降,与实测值作对比,结果较为接近。通过两个工程实例的应用,表明了本文提出的反分析算法是可行的,并且是一种通用算法,该算法考虑了目标实测值量测的误差和施工工况的影响,可以选择多个实测目标作为约束条件综合分析多个参数,提高了分析效率,可以用于指导设计施工,为岩土工程中土层参数反分析提供了一条有效途径。
[Abstract]:Soil layer parameter is an important index in the process of geotechnical engineering design. Generally, it is obtained by indoor test. Due to the natural variability of soil layer, test conditions, test method and other reasons, the test results are often discrete. Moreover, if the engineering construction method is different, even if the soil is the same soil, the soil layer parameters will be different. Therefore, it is necessary to adopt appropriate method to determine the soil layer parameters reasonably, combined with the existing monitoring data, The parameters of soil layer are obtained by means of back analysis, and then the parameters obtained by back analysis are used to carry out positive analysis, so that the data can be updated in time. To guide the design and construction of later stage or to guide similar engineering design. In this paper, the research work is carried out from the aspects of back-analysis method and its application, as follows: (1) in view of the problems existing in the present back-analysis method, A non-parametric regression based back-analysis algorithm for soil layer parameters is proposed. Taking the same measured target under different working conditions as the constraint condition and considering the error of the measured target measurement, the tolerance 蔚 is selected, and the tolerance range of the target measured value is obtained. Then the initial range of soil parameters is determined, and the response surface between soil parameters and measured targets is established by combining uniform test and non-parametric regression method, and a large number of parameter samples are obtained according to the enumeration of working conditions. A large number of target calculated values are obtained by using response surface interpolation. The calculated and measured values of the target are compared one by one, and the soil layer parameter data samples corresponding to the calculated value of the target in the tolerance range are sifted out. The statistical range of the data sample is taken as the range of actual values of soil parameters in a certain tolerance range. The application of inverse analysis algorithm in foundation pit engineering is discussed. The steel bearing force and the horizontal displacement of pile body are chosen as the measured targets. By using the inverse analysis algorithm, the shear strength indexes c and 蠁 of the soil layer in the depth range of the foundation pit profile supporting pile are back analyzed, and the interval of the soil layer parameters is obtained. The application of the back-analysis algorithm in shield tunnel engineering is verified. The anti-analysis algorithm is used to analyze the shear strength index (c, 蠁) of the soil layer in the influence area of shield tunnel excavation by selecting the ground settlement as the measured target. The value interval of soil parameters is obtained, and the ground subsidence near the excavation section is predicted by using the result of back analysis. The results are close to those of the measured values. Through the application of two engineering examples, The results show that the proposed inverse analysis algorithm is feasible, and it is a general algorithm, which takes into account the error of target measured value and the effect of construction condition. A number of measured targets can be selected as constraint conditions for comprehensive analysis of multiple parameters, which improves the efficiency of analysis, can be used to guide design and construction, and provides an effective way for back analysis of soil parameters in geotechnical engineering.
【学位授予单位】：昆明理工大学
【学位级别】：硕士
【学位授予年份】：2015
【分类号】：TU43

【参考文献】