The tests in this investigation were performed on a natural soft clay with plasticity index around 32%, which was K0 consolidated to a vertical stress of 2942 kPa and then K0 unloaded to a vertical stress of 74 kPa (i.e., to the “in situ” stress). The specimens so created were disturbed in various ways to study the effect of sample disturbance on the stress–strain relationships during undrained shearing and during drained K0 loading (i.e., K0 triaxial and oedometer tests). The results for two testing alternatives may be summarized as follows. Alternative 1: Allow the specimen to swell at the correct in situ effective stresses, but accept an initial water content that is higher than the in situ value. This alternative was found to give the best stress–strain relationships around the in situ effective stresses for undrained triaxial tests, but with undrained shear strength values up to about 20% too low, due to the swelling taking place during consolidation to the in situ effective stresses. Alternative 2: Prevent swelling by starting the test at effective stresses that are higher than the in situ stresses, but with a water content that is closer to the in situ value than if alternative 1 is chosen. Using only isotropic stresses prior to shearing, this alternative was found to give better undrained shear strength values (although up to about 14% too high) but strain values much too small around the in situ effective stresses. For oedometer tests, only alternative 2 was investigated. Also, for these tests, the strains around the in situ stress were too small, but preconsolidation stresses estimated from stress–strain curves were typically only around 60% of the true value.
A Method for Correcting Undrained Shear Strength for Sample Disturbance
