基于支持向量机的盾构施工地表沉降预警系统研究

发布时间：2019-04-27 16:40

【摘要】：我国已进入轨道交通建设高速发展期,如何更为有效地控制盾构施工引起的地表沉降、实现盾构施工地表沉降超前预警具有重要的现实意义。目前盾构施工技术领域正在积极开展有关操作更简便、预测更可靠及普适性更强预测方法的研究,为地表沉降预警提供更为可靠的解决方案。本文主要采用基于支持向量机的分析方法开展盾构施工地表沉降预警系统的研究。以广州市轨道交通四号线南延段某盾构区间为工程背景,在详细研究国内外盾构施工地表沉降预测技术的基础上,主要开展以下研究工作:(1)开展了有关上软下硬地层对盾构隧道地表沉降预测结果影响的研究。运用有限元数值模拟软件Midas GTS NX,建立考虑开挖面穿越复合地层和忽略开挖面穿越复合地层的两种三维有限元模型,对比分析其预测结果。进一步研究盾构推进过程中横向地表沉降的变化规律,并将模型预测结果与工程实测数据、Yoshikoshi经验公式法预测结果对比分析,用以检验三维数值模拟预测结果的可靠性。研究表明,当盾构开挖穿越复合地层时,应考虑复合地层分层情况;基于Midas GTS NX的上软下硬地层盾构法隧道地表沉降预测模型具有较高的精确度和可靠性,所输出预测值可较好地预测实际施工中的地表沉降。(2)在研究支持向量机理论和算法的基础上,将支持向量机技术与模糊信息粒化技术结合,运用于实现盾构施工地表沉降超前预警的研究工作中。研究由选择一种更为适用的支持向量机寻参方法开始,分别运用网格寻参法及粒子群优化寻参法,建立两种不同寻参方法的模糊信息粒化非线性回归预测分析支持向量机,对比分析其盾构施工地表沉降的预测结果。研究结果表明,在地表沉降预警系统中运用网格寻参法支持向量机更为可靠,可输出预测结果变化空间和变化趋势,是一种实现盾构隧道地表沉降超前预警的新途径。(3)在MATLAB软件图形用户界面(GUI)中,基于模糊信息粒化网格寻参法支持向量机,编写代码实现盾构施工地表沉降预警系统的建立。系统主要目标为针对不同地质条件自动提取储存的相关工程数据,实现数据查看与地表沉降预测,并完成对存在不安全因素盾构区间进行预警。为实现系统功能,从背景、数据读入、地表沉降预测、预警功能等方面开展了相关研究工作。开发的盾构施工地表沉降预警系统,有较好的可操作性,预测结果可靠度较高,对实际盾构隧道施工具有一定现实指导意义。
[Abstract]:China has entered the high-speed development period of rail transit construction. How to control the surface settlement caused by shield construction more effectively and realize the advance early warning of surface settlement in shield construction is of great practical significance. At present, the shield construction technology field is actively carrying out the research on the methods of more convenient operation, more reliable prediction and stronger universality, so as to provide a more reliable solution for the early warning of surface settlement. In this paper, the analysis method based on support vector machine (SVM) is used to study the ground settlement early warning system of shield construction. Based on the engineering background of a shield region in the southern extension of Guangzhou Rail Transit Line 4, based on the detailed research on the surface settlement prediction technology of shield construction at home and abroad, The main research work is as follows: (1) the influence of upper soft and lower hard strata on the prediction of ground subsidence of shield tunnel is studied. The finite element numerical simulation software Midas GTS NX, is used to establish two kinds of three-dimensional finite element models considering the cut face traversing the composite formation and neglecting the cut surface crossing the composite formation, and the prediction results are compared and analyzed. In order to verify the reliability of the prediction results of 3D numerical simulation, the variation rule of lateral surface settlement during shield propulsion is further studied, and the prediction results of the model are compared with the measured data, and the Yoshikoshi empirical formula method is used to predict the reliability of the prediction results. The results show that the delamination of composite strata should be considered when shield is excavated through composite strata. Based on Midas GTS NX, the prediction model of tunnel surface settlement with shield method in upper soft and lower hard strata has high accuracy and reliability. The output prediction can predict the ground subsidence in actual construction well. (2) on the basis of studying the theory and algorithm of support vector machine (SVM), the support vector machine (SVM) and fuzzy information granulation (FIS) are combined with each other. It is used in the research of advance early warning of ground settlement in shield construction. The research begins with choosing a more suitable parameter finding method of support vector machine, using grid parameter finding method and particle swarm optimization parameter finding method, respectively, to establish two different parameter seeking methods of fuzzy information granulation nonlinear regression prediction support vector machine. The prediction results of ground settlement in shield construction are compared and analyzed. The results show that the grid seeking parameter support vector machine is more reliable in the early warning system of surface subsidence, and it can output the change space and trend of prediction results. It is a new way to realize advance early warning of ground subsidence of shield tunnel. (3) in MATLAB software graphical user interface (GUI), based on fuzzy information graining grid search method, support vector machine (SVM), The code is written to realize the establishment of the ground settlement early warning system for shield construction. The main objective of the system is to automatically extract and store relevant engineering data according to different geological conditions, to realize data viewing and surface settlement prediction, and to complete the early warning of shield regions with unsafe factors. In order to realize the system function, the related research work is carried out from the aspects of background, data reading, surface subsidence prediction, early warning function and so on. The ground settlement pre-warning system developed for shield construction has good maneuverability and high reliability of prediction results, which has certain practical guiding significance for the actual shield tunnel construction.
【学位授予单位】：江西理工大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：U455.43

【参考文献】