Development of coupled centrifuge-numerical modelling to achieve a global tunnel-soil-structure interaction analysis

利用地下空间进行基础设施建设对现代城市的发展至关重要。英国伦敦横贯铁路项目的重大投资清楚地证明了这一点。隧道掘进已经发展成为一个复杂的施工过程，涉及使用自动化设备和专业材料。然而，即使使用现代方法，也无法避免因掘进而造成的地面变形。在现代城市中，由于地下基础设施丰富，地下空间有限，因此了解隧道开挖对附近地下基础设施和地上结构的影响非常重要。隧道开挖对地上构筑物的影响分析是一个极其复杂的土-结构相互作用问题。这个问题涉及隧道、土壤、地上结构的行为，以及土壤和结构部件之间发生的高度非线性相互作用。对个别领域（即孤立的土壤或结构）的分析可以提供总体行为的一些指示，但它不能给出全球系统的真实表示。现代设计和研究工具，如数值和岩土离心模型，有能力准确地单独模拟土壤或结构域，但很难复制涉及土壤-结构相互作用的大型复杂系统的全局行为。本项目将开发一种方法，其中岩土离心机和数值模拟技术以这样一种方式耦合在一起，以充分利用每种技术各自的优势，以便获得隧道建设相互作用问题的准确全局解决方案。该项目考虑了一个特定的场景，在一个建筑的桩基下建造隧道。采用诺丁汉大学土工离心机对隧道桩土基础区域进行了模拟。土工离心机允许在受控的实验室环境中测试全尺寸原型的小尺寸模型，并复制复杂的土壤和土壤-结构相互作用行为。离心机是必要的，以增加土壤的重量在小尺寸模型中，使地应力和行为在模型中符合那些在全尺寸原型；例如，离心机模型中0.1m深的一段土的重量与10m深的同一原型土的重量相同，当离心机在模型中旋转时，加速度场相当于100倍的地球重力。离心模型能比数值模型更真实地模拟隧道开挖引起的位移和由此产生的土-结构相互作用。对于隧道施工问题，采用数值模型求解基础施工区域。考虑了该领域的各种复杂性，如建筑-基础连接，建筑刚度和材料性能。该数值模型将提供对建筑物的精确模拟，这在按比例的离心机模型中很难实现。离心机和数值模型将通过数据接口系统进行耦合，该系统将在模型之间实时传递桩的位移和荷载信息。这种接口意味着离心机中的物理模型和数值模型之间的相互作用被捕获，并且正确地模拟了隧道-土壤-基础-建筑系统的整体行为。该项目的成功完成将代表着在改进研究隧道-结构相互作用问题的建模方法方面迈出了重要的一步，并将为帮助提高对这一复杂而重要的施工场景的理解提供有价值的信息。

The use of underground space for infrastructure development is vital for the growth of modern cities. This is clearly demonstrated by the significant investment in the Crossrail project in London, UK. Tunnelling has evolved into a sophisticated construction process involving the use of automated equipment and specialist materials. The use of modern methods does not, however, prevent the unavoidable result of ground deformations caused by tunnelling. In modern cities, where underground space is limited due to the abundance of buried infrastructure, it is important to understand the effects of tunnelling on nearby buried infrastructure and above-ground structures.The analysis of the effect of tunnelling on above-ground structures is an extremely complex soil-structure interaction problem. The problem involves the behaviour of the tunnel, the soil, the above-ground structure, and the highly non-linear interactions that occur between the soil and structural components. Analysis of individual domains (i.e. the soil or structure in isolation) can provide some indication of overall behaviour but it does not give a true representation of the global system. Modern design and research tools, such as numerical and geotechnical centrifuge modelling, have the ability to accurately model the soil or structural domains individually but struggle to replicate the global behaviour of large complex systems involving soil-structure interactions.This project will develop a method in which geotechnical centrifuge and numerical modelling techniques are coupled together in such a way as to take full advantage of the respective strengths of each technique in order to obtain an accurate global solution to the tunnel-building interaction problem. The project considers a specific scenario in which a tunnel is constructed beneath a building on a piled foundation. The University of Nottingham geotechnical centrifuge is used to model the tunnel-soil-piled foundation domain. The geotechnical centrifuge allows testing of small-scale models of full-scale prototypes within a controlled laboratory environment and replicates complex soil and soil-structure interaction behaviour. The centrifuge is necessary to increase the weight of the soil in the small-scale model so that ground stresses and behaviour in the model match those in the full-scale prototype; for example a 0.1m deep section of soil in the centrifuge model weighs the same as 10m of the same prototype soil when the centrifuge is spun to give an acceleration field equivalent to 100 times earth's gravity in the model. The centrifuge model can provide a more realistic simulation of the tunnelling induced displacements and resulting soil-structure interactions than a numerical model.For the tunnel-building problem, a numerical model is used to solve the foundation-building domain. Various complexities of this domain are considered, such as the building-foundation connections, building stiffness, and material behaviour. The numerical model will provide an accurate simulation of a building which would be very difficult to achieve in a scaled centrifuge model.The centrifuge and numerical models will be coupled through a data interfacing system which will pass information of pile displacements and loads between the models in real-time. This interfacing means that the interactions between the physical model in the centrifuge and the numerical model are captured and that the global tunnel-soil-foundation-building system behaviour is modelled correctly.The successful completion of this project will represent a significant step in improving modelling methods to study the tunnel-structure interaction problem and will provide valuable information to help improve the understanding of this complex and important construction scenario.