Effective Stress Failure Envelope Forms and Parameter Variability

Triaxial tests are often used to determine the effective stress strength envelope of soils. Although the linear Mohr-Coulomb envelope is often used in geotechnical engineering, the failure envelope can also be non-linear, with or without a shear stress intercept (i.e., cohesion). Selection of an appropriate mathematical form for the strength envelope is an important step in the interpretation of test data and characterization of soil properties. Thirty-three triaxial data sets were analyzed using a spreadsheet that determines best-fit envelopes for linear, two and three parameter power functions, secant ϕ', and modified Hoek Brown mathematical forms along with the coefficient of variation of the shear strength data. The trends in the type of envelope and variability are discussed. Variability in test data leads to uncertainty in the characterization of shear strength. Possible sources of such variability are the composition of the soil itself and error introduced during testing. The impact of the number of tests and variability on the linear model parameters is assessed for remolded clay and compacted clay. An increase in the number of tests reduces the uncertainty of linear model parameter values but has little effect on the predicted shear strength at a representative normal stress of one atmosphere.