某深基坑监测及变形预测模型研究

发布时间：2018-01-21 01:52

本文关键词： 深基坑工程灰色GM模型马尔可夫链 BP神经网络组合模型　出处：《武汉理工大学》2013年硕士论文　论文类型：学位论文

【摘要】：随着我国经济的高速发展,基建投资的加大,深基坑工程的安全成了建筑安全的重点,其支护形式、安全监测与变形预测则就成了岩土工程领域的重要研究课题。基坑在开挖过程中,必然会引起基坑的变形。然而,如何找到一种科学准确的基坑形变预测方法对工程实践具有重要意义。本文以武汉王家墩中央商务区地下交通环廊第一标深基坑工程为背景,分析了地下交通环廊的SMW工法桩、双轴搅拌桩及高压旋喷桩支护的施工方法,分析了地下交通环廊深基坑的特点。在前人的研究基础上,分析了灰色系统理论和马尔可夫理论及BP神经网络原理,并建立了基于灰色系统、灰色马尔可夫链、BP神经网络及灰色马尔可夫-BP神经网络的变形预测模型。最后以王家墩中央商务区地下交通环廊第一标深基坑为实例进行预测,在MATLAB7.0的工具箱上编译四种理论方法,通过选择不同的参数训练,经对比分析说明BP神经网络模型用于深基坑变形预测精度较其它模型要高,能很好的预测未来变形的发展趋势,而GM(1,1)模型、灰色马尔可夫模型及灰色马尔可夫-BP神经网络模型在数据不多的情况下可以作为一种变形预测手段。本文得出了以下结论： 1)由于BP神经网络自身的容错能力和自适应学习能力,本文在地下交通环廊深基坑监测预测中,其模型的预测精度高于灰色GM(1,1)模型和灰色马尔可夫模型。 2)结合灰色GM(1,1),马尔可夫链、BP神经网络各自的优点,建立了灰色马尔可夫-BP神经网络预测模型。在面对地下交通环廊深基坑的监测预测中,少量样本信息也可以获得较高的精度。 3)利用所建立的模型对武汉王家墩中央商务区地下交通环廊第一标深基坑工程在开挖过程进行监测中,对基坑围护结构变形进行了预测,通过与实测值的对比研究,表明本文所建立的预测模型有较好的适用性。
[Abstract]:With the rapid development of our economy and the increase of capital investment, the safety of deep foundation pit has become the key point of construction safety, and its supporting form. Safety monitoring and deformation prediction have become an important research topic in the field of geotechnical engineering. How to find a scientific and accurate method of foundation pit deformation prediction is of great significance to engineering practice. Based on the first standard deep foundation pit project of the underground traffic ring corridor in Wuhan Wangjiadun central business district, this paper analyzes the construction methods of the SMW pile, the two-axis mixing pile and the high-pressure rotary jet pile supporting the underground traffic ring corridor. The characteristics of deep foundation pit of underground traffic ring corridor are analyzed. On the basis of previous research, grey system theory, Markov theory and BP neural network theory are analyzed, and grey system based on grey system is established. The deformation prediction model of grey Markov chain BP neural network and grey Markov BP neural network is presented. Finally, the first standard deep foundation pit of underground traffic ring corridor in Wangjiadun central business district is taken as an example. Four theoretical methods are compiled on the toolbox of MATLAB7.0. By selecting different parameter training, it is proved that the BP neural network model is more accurate than other models in predicting the deformation of deep foundation pit. It can well predict the future trend of deformation, and GM1 / 1) model. The grey Markov model and the grey Markov BP neural network model can be used as a method to predict the deformation when the data are not much. In this paper, the following conclusions are drawn: 1) because of the fault tolerance and adaptive learning ability of BP neural network, the prediction accuracy of the model is higher than that of grey GM(1 in the monitoring and prediction of deep foundation pit of underground traffic ring-corridor. 1) model and grey Markov model. 2) the advantages of Markov chain BP neural network combined with grey GM1 / 1 / 1 / 1 / 1 / 1 / 1 / 1 / 1, respectively. The grey Markov BP neural network prediction model is established. In the monitoring and prediction of deep foundation pit of underground traffic ring corridor, a small amount of sample information can also obtain high accuracy. 3) using the established model, the deformation of the retaining structure of the foundation pit is forecasted during the excavation process of the first standard deep foundation pit of the underground traffic ring corridor in the central commercial district of Wangjiadun, Wuhan. The comparison with the measured data shows that the proposed prediction model has good applicability.
【学位授予单位】：武汉理工大学
【学位级别】：硕士
【学位授予年份】：2013
【分类号】：TU753

【参考文献】