Factors controlling the stability of compacted clay embankments

Abstract

Factors affecting the stability in clay fill embankments in the Auckland region have been investigated by studying two types of Waitemata residual clay from Auckland: a clay of high plasticity and a silty clay of low plasticity. The tests showed that the undrained shear strength (Su) on the soil compacted in the standard compaction test reduced rapidly with increase of water content. The value of the shear strength depended on the test method used, the hand vane giving about 40% greater values than the unconfined compression test. The relationship of the Su to the water content, air voids and pore water pressure in the soils were established. Su increased with reduction of air voids. At a confining pressure, pore water pressure increased with increase of water content and reduction of air voids. It is noted that compacting fill to high air voids may avoid the development of high pore water pressure in high embankments. The effective shear strength parameters, c’ and ɸ’, were obtained for the two soils showing that they were not noticeably affected by water content and dry density. The analysis of idealised embankments showed that safety factors from the total stress method were greater, and the critical surfaces deeper than those from the effective stress method. The two methods assume different stress path with the effective stress. ignoring the dilatant nature of the compacted fill and this was explained with examples. It is suggested that values from the hand vane test should be used with caution as they give large safety factors. Special stress path test done showed that if dilating soil samples being sheared in undrained condition, were changed to drained by providing a source of water at constant pressure, they would attract in water and the shear strength of the samples would be reduced. This suggested that the Su from the unconfined compression test might not be attained if the dilating critical zone in an embankment attracted water.