CHAPTER 1

INTRODUCTION

1.1 Background

The scarcity of land in Singapore requires maximum utilisation of underground space. At present, basement car parks and shopping malls are common and strategic projects such as the construction of an underground road system has been planned in the near future. As a result, excavation works associated with these types of construction are becoming more and more common (Tan et al., 1995) in Singapore.

In urban areas, excavation works usually utilise vertical cutting technique with diaphragm wall or sheet piles braced with cross struts. The top-down construction technique with diaphragm wall is also practised in Singapore. Generally, excavation causes the reduction of horizontal pressure on the side of excavation, leading to the movement of soil behind the retaining walls towards the cut. Meanwhile, the relief of vertical pressure in the soil may cause heave to occur at the base of an excavation. The bottom heave and the inward movement of soils are often accompanied by subsidence of the ground surface surrounding the cut. If the excavation is carried out below the ground water level and the permeability of the soil layer is relatively high, the seepage-induced consolidation and draw-down may also cause problems (Lee et al., 1993). Due to these phenomena in excavation works, excessive soil deformation can cause serious damages to retaining structures and adjacent structures unless proper earth retaining and support systems are designed and employed.

In Singapore, most of the deep excavation works are situated in densely built-up areas. In many areas, the soil deposit is the Marine Member of the Kallang Formation. The Marine Member is a deposit of normally consolidated homogeneous blue grey marine clay with a shear strength of less than 25 kPa. This type of clay has caused stability problem to the retaining system. Moreover, it can also induce significant ground movements around the excavation.

The effects of nearby excavation on existing structures such as historical buildings, tall buildings with deep foundations, underground tunnels and major utilities tunnels as shown in Fig 1.1, are increasingly a concern for engineers. When excavation work is carried out close to existing Mass Rapid Transit (MRT) tunnels or stations, the MRT Code of Practice for Railway Protection (1996) states that "differential movement resulting from the works shall not produce a final distortion in the plinth or track in excess of 3 mm in 6 m or total movement in the MRT structure or tracks exceeding 15 mm in any plane". As this stringent requirement must be followed, the associated ground movements during excavation need to be monitored closely.

The estimation of the ground movements caused by excavation is becoming important in the presence of nearby critical structures. Traditionally, safety factors are used in various ways to indirectly control the displacement (Simpson, 1992). However, such an approach will not be able to give a precise estimate of the displacement (Tan et al., 1995). For a site with a particular geometry and supporting system, the influence of construction sequence, as well as the rate of dissipation of excessive pore pressure will also affect the profile of ground movement. However, many practising engineers in Singapore are still using traditional empirical methods to estimate ground movements as these are simple and easy to use. Unfortunately, these methods do not account for some of the important factors stated above.

For excavation work around a critical structure, numerical analysis has provided a more rational approach for evaluating ground movement. With appropriate soil models, numerical method can be employed. Though the real problem is three-dimensional, a two-dimensional plane strain problem is usually analysed as it involves significantly less efforts. The properties to be used in the models are estimated from site investigations, laboratory tests as well as correlated results obtained from experience. Generally, the deflection of the wall can be estimated reasonably well with appropriate selection of these parameters. However, Tan et al. (1995) pointed out that the ability of predicting displacement at some distances away from the excavation has been less successful. Results from numerical analysis could not explain for the more abrupt decrease of the settlement with distance that is usually observed in the field.

The accuracy of predicting movement behind the retaining wall has important implications to engineers in their strive for a more economical design. Thus, it is important to understand clearly various controlling factors relating to such movements. In recent years, there is an increasing awareness that the stiffness at very low strain is actually much higher than the stiffness obtained from conventional triaxial testing. On the other end, with large deformation, the continuum approach of finite element method is not able to simulate the development of shear plane. Since existing field data are not able to provide a clear understanding of these problems, it is important to conduct controlled experiments. This will provide a better understanding of the mechanism of excavation. Centrifuge models tested in a high-g environment can be used to simulate the prototype behaviours of excavation in order to investigate the importance of various controlling factors related to this problem. This technique ensures that the stress and strain regime is correctly reproduced in the model, allowing a realistic simulation of the mechanics of such an excavation in a prototype.