Centrifuge modelling of tunnelling close to a diaphragm wall

The experimental set-up and some results of a series of centrifuge tests which were performed at City University are shown and discussed. Tests were carried out on a circular model tunnel in a clay ground and in undrained conditions. The basic idea being tested was whether or not a diaphragm wall embedded at different locations close to a shallow tunnel could reduce movements in the ground extending beyond the wall and further away from the tunnel. The length of the embedded diaphragm wall was varied, as well as its thickness. The influence of the soil-wall interface was also studied. An image processing system was used to measure displacement fields in the models. The influence of the tunnel support pressure on the displacement field is discussed. Also, the influence of the walls edge and of the embedding technique is shown. In particular two different ways of achieving a rough interface are compared. A significant influence of the wall self-weight and of the interface roughness on the model stability was observed during the tests. Two kinematically admissible collapse mechanisms are analysed in order to understand these experimental findings.