Project Overview
When a new metro station or a deep basement are to be constructed in a city, a large hole in the ground is needed. The hole needs to be safe to work in and to allow access and very often the chosen solution is to support the sides of the large hole with a braced embedded retaining wall. These are substantial pieces of temporary works of considerable cost. Recent examples are the 230 long by 24 m wide by 23m deep excavation for the new Crossrail station at Paddington in London. An important feature of this form of construction is large props that span between the walls, to hold them up.
For those tasked with designing the prop size, location, number and the walls the key issues are prediction of ground movements adjacent to the excavation (which could negatively affect buildings) and propping forces (so that the right props can be used and while guidance exists for designers in some recent publications produced by the UK construction information organisation, CIRIA, coverage of the behaviour at excavation corners as regards both design issues is poor. There is substantial published research on the computational modelling of braced excavations but only in two-dimensions (i.e. s slice through a long wall), some of it validated against field data, however accounting for 3D effects as required for the analysis of corners is rare and insubstantial. Improving our understanding of the behaviour of these corners and how it is affected by soil behaviour, system stiffness, and prop loading will lead to (a) greater economy in propping schemes and (b) more certainty in the prediction of ground movements adjacent to corners, potentially reducing the accommodation works required to prevent damage to adjacent structures.
The programme of research proposed here comprises complex computational simulations of the construction of a braced excavation, taking into account differences in geometry, materials and sequences. The problem can only be properly tackled using a 3D model (unlike many other problems in geotechnical engineering) however even today, the computational tools we use struggle to deliver results quickly when we try to model in 3D. So, in this proposal we will be using a clever method where "reduced order models", (ROMs) will be made, using results from a relatively small set of the complex 3D models. A ROM is much easier to use and is generated by manipulation of a limited number of the high fidelity 3D simulations. From these ROMs we will derive results and prepare guidance for engineers designing braced excavations which will enable cheaper and simpler schemes to be used.
The aim of this project is to develop new and more accurate ways to predict ground movements and prop loads for large braced excavations as regularly used around the world for the construction of new underground transportation infrastructure. The objectives to deliver this vision are as follows:
- to generate a library of data from 3D finite element (FE) analyses of models of corners of braced excavations;
- to undertake a programme of centrifuge modelling of corners of braced excavations for a selection of key scenarios to validate the 3D modelling approach used in (a);
- to validate the FE analyses against field data;
- to use the results from (a) to build reduced order models (ROMs) of the braced excavation corner problem;
- to use the ROMs to explore the problem at high efficiency; and
- to produce a design guide to supplement current advice and to disseminate this to practitioners.
