深厚人工填土路基的沉降计算与预测研究

发布时间：2018-04-22 21:15

本文选题：人工填土 + 沉降计算　；参考：《湘潭大学》2017年硕士论文

【摘要】：由于自然环境和地理条件的影响,我国的西部,特别是云贵地区,地势起伏较大,地形地貌复杂多变,导致深厚人工填土路基十分常见,而沉降是影响其工程质量的一个关键问题。在实际工程中影响路基沉降的因素众多,而各因素之间的变化规律、相互联系以及对沉降规律的认知程度等问题,往往难以分析清楚,使得沉降量的理论计算和工后沉降量的预测均存在一定的局限性,因此准确的计算与预测路基沉降就具有重要的理论价值和工程意义。本文依托《中石油云南石化铁路专用线厂区铁路装卸场路基沉降观测》项目,结合现场大量沉降观测数据,采用理论计算和模型预测对中石油云南石化铁路专用线厂区铁路装卸场深厚人工填土路基的沉降进行了研究,取得了以下主要成果:(1)针对中石油云南石化铁路专用线厂区铁路装卸场深厚人工填土路基,对其进行了沉降计算,并对比分析了不同沉降计算方法对深厚人工填土路基的适用性。根据现场实际情况,选取典型工点进行钻探取样,考虑人工填土的特点,开展了人工填土基本物理工程特性试验和单向压缩固结试验,并获取人工填土的物理参数、压缩系数、压缩指数等压缩指标,进行了沉降理论计算,并对比分析不同沉降理论计算方法对深厚人工填土路基沉降理论计算的适用性。结果表明:规范修正公式法和应力历史法法对人工填土路基最终沉降具有一定的指导意义。(2)针对不确定性影响因素下实测沉降数据工后沉降预测误差较大的问题,采用递归图和递归定量分析方法对沉降时间序列的可预测性进行了分析,辨识出预测效果较好的沉降时间序列。通过对深厚人工填土路基的沉降时间序列进行可预测性分析得出:沉降时间序列具有混沌特性;其的递归图颜色分布越规律、色彩深浅变化越均匀,沿着主对角线出现的白色区域越小,表明可预测性越强;通过提取递归定量分析指标随时间变化的曲线,直接确定最佳预测时间起点为22天;利用递归图求取二阶任意熵作为沉降时间序列的可预测步长因子,推算出C4沉降时间序列的预测效果最佳。(3)利用残差马尔科夫链修正无偏灰色预测模型对深厚人工填土路基的沉降进行了预测,获得的拟合曲线与实测曲线吻合度较好,验证了该模型的准确性。在实际的监测过程中,所获取的沉降时间序列是非等时距的,对这一类的数据进行预测,首先采用工程上常用的三次样条插值函数进行等时距处理,然后采用残差马尔科夫链修正无偏灰色模型进行沉降预测,其预测精度达到要求。
[Abstract]:Because of the influence of natural environment and geographical condition, the western part of our country, especially in Yungui area, has a large relief and complex landform, which leads to the deep artificial filling roadbed very common. Settlement is a key problem affecting the engineering quality. There are many factors that affect subgrade settlement in practical engineering, but it is difficult to analyze the law of variation, interrelation and cognitive degree of settlement law among the factors. The theoretical calculation of settlement and the prediction of settlement after construction have some limitations, so it is of great theoretical value and engineering significance to calculate and predict the settlement of roadbed accurately. Based on the project of "observation of subgrade settlement of Railway loading and unloading Yard in the Plant area of Yunnan Petrochemical Railway of PetroChina", this paper combines with a large number of on-site settlement observation data. Theoretical calculation and model prediction are used to study the settlement of the deep artificial subgrade in the railway loading and unloading yard in the special railway line area of CNPC Yunnan Petrochemical Company. The following main achievements have been obtained: (1) the settlement calculation has been carried out for the deep artificial filling subgrade of the railway loading and unloading yard in the special railway line area of CNPC Yunnan Petrochemical Company. The applicability of different settlement calculation methods to deep artificial embankment is compared and analyzed. According to the actual situation of the site, the typical work points are selected for drilling sampling, considering the characteristics of artificial fill, the basic physical engineering characteristic test and unidirectional compression consolidation test of artificial fill are carried out, and the physical parameters and compression coefficient of artificial fill are obtained. The compressing index is used to calculate the settlement theory, and the applicability of different settlement theory calculation methods to the deep artificial subgrade settlement calculation is compared and analyzed. The results show that the modified formula method and the stress history method have certain guiding significance for the final settlement of artificial fill subgrade. The prediction of settlement time series is analyzed by using recursive graph and recursive quantitative analysis method, and the settlement time series with better prediction effect is identified. Through the predictive analysis of the settlement time series of deep artificial fill subgrade, it is concluded that the time series of settlement has chaotic characteristics, the more regular the color distribution of the recursion graph, the more uniform the change of color depth and light is. The smaller the white area along the main diagonal is, the stronger the predictability is, and by extracting the curve of recursive quantitative analysis index changing with time, the best prediction time starting point is 22 days directly. The second order entropy is obtained by recursive graph as the predictable step factor of settlement time series. The best prediction effect of C4 settlement time series is deduced.) the modified unbiased grey prediction model of residual Markov chain is used to predict the settlement of deep artificial fill subgrade. The fitting curve is in good agreement with the measured curve. The accuracy of the model is verified. In the actual monitoring process, the settlement time series obtained are non-iso-time interval. To predict this kind of data, the cubic spline interpolation function commonly used in engineering is first used to process the equal time interval. Then the residual Markov chain modified unbiased grey model is used to forecast the settlement, and the prediction accuracy is up to the requirement.
【学位授予单位】：湘潭大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：U416.1

【参考文献】