The application of critical state soil mechanics to predict ground deformations below an embankment constructed on soft alluvium
An embankment has been constructed to a maximum height of approximately twelve metres at Beckton, east London. The embankment is underlain by up to eight metres of alluvium and made ground. The embankment was monitored during construction. Centreline settlements of the order of O.6m and horizontal movements at the embankment toe of up to O.1Sm were measured. The excess pore pressures generated during construction were found to dissipate rapidly. Soil samples recovered from the site have been tested at city University. The alluvial clay is shown to have extremely variable compression characteristics and it is not possible to determine a single set of critical state soil parameters from the "undisturbed" soil samples. The use in design of parameters derived from reconstituted soil samples is investigated. Laboratory test results indicate that a small quantity'of organic matter in the soil has a large effect on the engineering properties of the alluvial clay. The importance of not dry~ng soils which contain organic matter is highlighted. It is shown that, despite the effects of the organic matter, critical state concepts and in particular the modified Cam clay soil model can be used to predict the behaviour of the alluvial clay in a wide variety of laboratory tests. standard methods of determining the yield stress or preconsolidation pressure are shown to be inadequate when used to analyse the results of compression tests performed at a constant rate of loading. An alternative method, the Lambda construction, is proposed and this is shown to be both reliable and objective. Two distinct testing methods have been used to probe the state boundary surface of the reconstituted alluvial clay samples. The results of these tests appear to indicate the existence of a single state boundary surface for both Ko and isotropically compressed samples. This surface has a shape similar to that predicted by the modified Cam clay soil model. This result conflicts with the findings of many other investigators. Finite element modelling of the embankment construction has been carried out. A single set of soil parameters derived from tests on the reconstituted clay were used to represent the alluvial stratum. The construction of the embankment was modelled as a drained event. The results are shown to compare favourably with the actual field data.