基坑开挖的数值模拟及变形预测研究-岩土工程专业论文.docxVIP

摘 要 随着城市建设的发展和人民生活水平的提高，多层及高层建筑物的地下室、地 下车库、地铁、大型桥梁等在建设过程中都会面临不同程度的深基坑工程问题。特 别是在基坑开挖的过程中桩顶的水平位移与基坑周围地表的沉降问题一直是被 密 切关注的，也是亟待解决的。由于岩土工程的受力状态和地质参数是复杂的、不确 定的、非线性的，因此需要找出一个在基坑开挖过程中能够实时提供参考依据的模 型，以达到施工设计、措施以及保障施工安全的目的。 本文从工程实际出发，以沈阳恒隆市府广场为例，对基坑现场监测的变形数据 进行分析处理，结合数值模拟与有限差分模拟对基坑的维护结构及地表沉降进行分 析预测，并融合了有限差分分析、时间序列、神经网络、遗传算法、MATLAB 等多 项理论与工具。其中，应用有限差分模拟了基坑的开挖过程，合理地给出了在基坑 开挖过程中的各应力分析云图以及桩顶变形云图，完成了施工过程中的有限差分模 拟的过程。应用时间序列和遗传优化神经网络对桩顶位移与周围地表沉降数据作了 进一步地分析预测处理，进而建立了时间序列和遗传优化神经网络模型，拟此预测 基坑及其地表沉降的变形值，使其能够为施工过程提供一个实时预测分析的参考依 据，防止施工环境的破坏，以起到提前做好应对措施的作用。本文最后把时间序列 模型与遗传优化神经网络封装成为一个系统，作为人机交互的平台，为今后类似的 工程提供了某种程度上的便利。 关键词：有限差分；时间序列；遗传算法；神经网络； - I - Abstract With the development of urban construction and peoples living standards, multi-storey and high-rise buildings in the basement, underground garage, subway, large-scale bridges, etc. in the construction process will be faced with varying degrees of deep foundation engineering problems. Especially in the process of excavation the level of pile top displacement and excavation of the settlement issue, which has been around the surface was closely, and it is urgent to solve. As the stress state of geotechnical engineering and geological parameters is complex, uncertain and nonlinear, therefore, we need to find one in the excavation process to provide real-time reference model in order to achieve construction design, the measures and the protection of construction safety purposes. This article is from engineering reality. It took Shenyang Hang Lung Plaza City Hall for example, pit-site monitoring of the deformation analysis of the data processing, combined with numerical simulation and finite-difference simulation of the excavation to maintain the structure and surface subsidence analysis and forecasting, and integration of the finite difference analysis, time series, neural networks, genetic algorithms, MATLAB, etc. a number of theories and tools. Among them, application of finite-differenc