The influence of spatial variability on the geotechnical design properties of a stiff, overconsolidated clay.

Abstract

The research presented in this thesis focuses on the spatial variability of the Keswick and Hindmarsh Clays within the Adelaide city area. Keswick Clay is locally significant since many of Adelaide's multi-storey buildings are founded directly on it, and internationally significant, since it has been shown by Cox (1970), that this clay exhibits remarkably similar properties to those of the well-documented London Clay. The assessment of the small-scale variability of the undrained shear strength of these clays is based on measurements obtained using the electrical cone penetration test (CPT), and a micro-computer based data acquisition system, designed specifically for this study. A significant feature of the data acquisition system is that it enables measurements to be obtained at intervals of 5 mm, both reliably and efficiently. The development of the data acquisition system is discussed, and the accuracy of its measurements is examined. The small-scale variability of the undrained shear strength of the Keswick Clay is based on more than 200 vertical CPTs, performed within an area of 50 X 50 metres at a site located in the Adelaide city area. The CPTs were spaced at lateral intervals varying between 0.5 and 5 metres, with each vertical CPT extending to a typical depth of 5 metres. In addition, the small-scale horizontal spatial variability of the Keswick Clay is examined using an electrical cone penetrometer driven horizontally into the face of an embankment, again located within the Adelaide city area. The accuracy of the CPT measurements is examined, and discussion is given of the shortcomings associated with a commonly used technique, by Baecher (1982), for estimating the random measurement error associated with various test procedures. The assessment of the large-scale spatial variability of the undrained shear strength of the Keswick and Hindmarsh Clays is founded on a data base of geotechnical engineering properties, compiled from a number of consulting engineering practices and government instrumentalities. The data base, known as KESWICK, contains approximately 160 site investigations, 380 boreholes, and 10,140 measurements obtained from a number of different laboratory and in situ tests. In addition, KESWICK is used to establish generalised trends and bounds, associated with the various geotechnical engineering design properties contained within the data base. The techniques of random field theory and geostatistics are used to quantify, model and predict the spatial variability of the Keswick and Hindmarsh Clays. These techniques are compared with one another in order to assess the suitability and shortcomings of each, when applied to the study of the spatial variability of geotechnical engineering materials. Furthermore, a number of specifically-written computer programs, which were developed to enable the various spatial variability analyses to be performed, are discussed. It is demonstrated that the lateral undrained shear strength of the Keswick Clay, within the Adelaide city area, exhibits a nested structure; that is, one which is the compound effect of several genetic sources of spatial variation. In addition, it is shown that this nested structure can be adequately modelled by means of a spherical semivariogram model. The nested structure is used, together with the kriging estimation process, to provide preliminary estimates of the undrained shear strength of the Keswick Clay, within the Adelaide city area. The analyses demonstrate that the nested model and the kriging process provide a useful facility for generating preliminary estimates of the strength of the clay. Finally, the significance of the spatial variability of the undrained shear strength of clay soils is examined, with reference to the design of embankments and pile foundations. It is demonstrated that the correlation distance can greatly influence the design of each of these geotechnical systems.