Effects of Shear Band Formation in Triaxial Extension Tests

The stress-strain and strength behavior from conventional triaxial extension tests is usually erratic due to excessive influence of necking in the hardening regime followed by shear banding. Two techniques have been attempted to impede the onset and gross development of strain localization. These resulted in macroscopically more uniform strains in the hardening regime and therefore more reliable stress-strain and strength behavior in triaxial extension. The results of tests on three different sands are compared with those from conventional extension tests in which strain localization can progress due to the soft rubber membrane. The effectiveness of each of the techniques is evaluated in view of its ability to maintain uniform strains and in view of the resulting sand behavior. The influence of the type of results on the choice of 3-D failure criterion is discussed.