Abstract
Different construction sequences may be used in braced excavations. In the conventional or “ordinary method”, temporary retaining walls are initially placed and struts are used to brace them as the excavation proceeds downwards. In an “alternative method”, sometimes called “inverted excavation”, the tunnel walls are used to retain the surrounding earth, while the roof slab helps to brace the excavation. Laboratory model tests were devised to simulate these construction procedures. The tests were carried out under two-dimensional conditions, using aluminum rods to represent the soil mass. The walls were constructed with aluminum plates fully instrumented with strain gauges in both sides to measure the bending strains of the wall. Cylindrical bars with and without springs were used to brace the excavation in the ordinary method; while an instrumented aluminum block was used as the top slab in the alternative method. As excavation proceeded, all data relative to surface settlements, wall deflections and bending moments, as well as axial strut loads was carefully recorded. Later the laboratory tests were simulated with finite element analyses using the recently proposed subloading tij model (Nakai and Hinokio, 2004). The numerical results were compared with the experimental data obtained during the models tests for both construction methods. The overall recorded behavior in terms of displacements, deflection, bending moment and axial load could be reproduced with striking accuracy both qualitatively and quantitatively. This shows the capability of the model to represent the complex behavior of granular materials even under the low stress range used in the model tests. The results of model tests and numerical analyses show that the alternative method is viable and effective in controlling induced surface settlements, provided that the tunnel walls are constructed with an appropriate thickness.
