The buttress wall is an alternative construction method for protection of adjacent buildings during excavation. It is generally recognised, but not verified, that the method has good effects in reducing wall movement and ground settlement, because buttress walls behave as T-shaped concrete beams, which possess comparatively larger stiffness than two-dimensional diaphragm walls. However, the behaviour of buttress walls in excavations is not fully understood, because of a lack of well-monitored case histories and systematic studies. In this study, the effectiveness of buttress walls in reducing wall deflection and ground settlement is evaluated through observed behaviour of a case history, and through finite element studies. The case study indicates that the use of buttress walls in a deep excavation can significantly reduce the deflection of a diaphragm wall. Effective prediction of the behaviour of the buttress wall can be achieved using the three-dimensional finite element method. Parametric studies reveal that deflection of the diaphragm wall is strongly influenced by the condition of the restraint at the bottom of the buttress walls, and by the number of buttress walls. As long as the buttress wall bottom is well restrained, and reasonable numbers of buttress walls are installed, wall deflection can be considerably reduced.

A finite element analysis of a hypothetical excavation was conducted to investigate recent stress-history effects on excavation-induced ground movements. The numerical simulations were conducted using a hypoplasticity model for clays which is capable of modeling small strain nonlinearity, soil anisotropy, and recent stress-history effects. The recent stress-history effects noted in previous laboratory experiments were reasonably reproduced by the model. Numerical simulations of the hypothetical excavation made through glacial clays are presented to illustrate the effect of preexcavation recent stress-history on computed ground movements. Lateral wall deformations caused by excavations in these soils with different recent stress histories resulted in computed movements that differed by a factor of two. These results suggest that recent stress-history effects should be considered when serviceability requirements dictate the design of a support system for which deformations must be limited to prevent damage to adjacent structures or infrastructure.