Perhaps the most common problem in engineering design involving soils, is the problem of calculating the movements that result from application of general states of stress to the soils. Until the last ten to twenty years, it was feasible to calculate either the very small movements that resulted from small loads (compared to the load that causes collapse), or to calculate collapse loads and assume that movements were so large that it became unnecess- ary to calculate them. For common cases in which neither extreme was valid, it was necessary to apply "engineering judgement." With the advent of modern digital computers, researchers are engaged in a two-pronged attack on this class of problem. One mode involves attempts to measure generalized soil properties (complete state of strain in response to a generalized state of applied stress) and development of models to describe the soil response. The second mode involves development of numerical methods so the resulting soil properties can be used to predict mass movements in the field. This proposal involves the first mode. The goal of this research is to perform reliable measurements of the response of samples of clean sands to complex states of stress, and then to fit appropriate models to allow analytic description of the resulting properties. The socalled hollow-cylinder device will be used to measure properties. A variety of constitutive models will be tried. Both monotonic loading (load directly to failure) and cyclic loading (repeated application of stresses less than the failure stresses) will be used.
