By Sarah Stallebrass, reseach student,Geotechnical Engineering Research Centre, City University. This paper was first published in GE‘sJanuary 1990 edition
This paper, which was awarded the Cooling prise for 1989, sets outs factors which should be tsjzen into account when investigating the small strain stiffness ofsoils in order to arrive at a realfistic prediction ofground movement.
At small strains, the behaviour of overconsolidated soil is non-linear and dependent on the current state and stress histozy ofthe soil. In addition to the overall history ofthe soil, described by the overconsolidation ratio, the recent stress history, for example a sudden change in direction of loading, has a major effect on the subsequent small strain soil stiffness (Som, 1968; Atkinson, 19732; Richardson, 1983).For realistic predictions of ground movements stress-strain relationships should be used which correspond not only to the correct state, stress history and strain level but also reflect the appropziate recent stress history. This is particularly important on sites where the recent stress history of the soil may change significantly across the site due to local geological variations or to nearby construction. Methods of modelling this non-linear, recent history dependent behaviour include empizical models derived from specific experimental data (Jardine and Potts, 1988),theoretically based constitutive soil models which only require basic soil parameters (Mroz et al, 1979) or mixed models which combine theory and empirical formulations (Simpson et at, 1979).