Estimating critical state soil mechanics parameters from shear box tests

Summary Prediction of compaction, tillage, root growth or other soil deformation events requires a description of the stress–strain properties of the soil such as the critical state model, but estimating the parameters is time consuming and expensive. I have developed a method of estimating critical state properties from a single shear box test, both saving much labour and providing more information than traditional analyses. The method is based on critical state analyses of the constant stress and constant volume shear box tests using the total stresses applied at the boundary. It derives the critical state property parameters from test data by minimizing the difference between test data and the simulated soil deformation (and hence the properties used in that simulation). The minimization is a form of regression analysis. The analyses resulted in good simulations of the history of states in space defined by τ xy γ xy and e space (constant stress test) or τ xy γ xy and σ y space (constant volume test). For normally consolidated samples, the analysis of a single constant stress test provided estimates of the slope of the critical state line ( M ), the slope of the virgin compression line (λ), the slope of the rebound line ( K ), and the elastic modulus ( E ). The standard deviation of the estimate of k could not be found. By contrast, the analysis of the constant volume test resulted in poor estimates, particularly of λ. This is because the test yields no information on changes of volume during deformation: hence volume change parameters are not successfully estimated. The constant volume test is therefore not suitable for back analysis.