Master of Science / Civil Engineering / Dunja Peric / The conventional triaxial test is the primary laboratory test for determining the shear strength of soils. Geotechnical field conditions such as long earth dams, long embankments, long retaining walls, strip foundations, tunnels, and buried pipelines often experience plane strain states of stress. However, stress strain and load deformation responses in plane strain loading differ considerably from responses observed in the conventional triaxial test. Research has shown that soils loaded in a plane strain state are far more sensitive to imperfections than soils tested in a conventional triaxial device. Plane strain loading leads to material instability manifested as sudden localized failure, resulting in decreased load-carrying capacity of the soil and compromised geotechnical and civil infrastructures. Although previous studies have mostly focused on granular materials, this research investigated the plane strain response of clay. An undrained plane strain compression test program was devised to investigate the effects of past stress history and strain rates on strain localization in kaolin clay. Experiments were carried out in a plane strain (or biaxial) device at Northwestern University, Evanston, Illinois. Because the device was heavily internally instrumented, strain localization progress was closely monitored throughout each biaxial test. Clay response in the biaxial test demonstrated three phases: (1) a homogenous response, (2) the onset and propagation of strain localization, and (3) the evolution of strain localization as a shear band. The duration of each phase was determined for each test, and a Lagrange strain tensor was used to obtain the evolution of volumetric and shear strains at the level of a shear band for three tests. Results revealed the development of large strains in these mesoscale structures. Furthermore, evolution of Mohr-Coulomb effective shear strength parameters was traced throughout the propagation and evolution phases by using two different methods. It showed that in clay samples, unlike in granular materials, the post-peak plateau, which is reached by deviatoric stress, corresponds to friction values that are significantly lower than the critical state values. Other researchers who used scanning electron microscope and anisotropy of magnetic susceptibility detected a significant reorientation of clay particles inside shear bands. Their findings combined with findings in this study lead to the conclusion that the sub-meso scale mechanism responsible for large shear strains and a severe reduction in effective friction is a significant reorientation of clay particles inside shear band.

Past studies on sands have shown that the shear strength measured in plane strain tests was higher than that measured in triaxial tests. It was observed that this difference changed with the friction angle &amp / #966 / cv at constant volume related to the mineralogical composition. In order to investigate the difference in strength measured in the wedge shear test, which approaches the plane strain condition, in the triaxial test, and in the shear box test, Anatolian sands were obtained from different locations in Turkey. Mineralogical analyses, identification tests, wedge shear tests (cylindrical wedge shear tests (cylwests) and prismatic wedge shear tests (priswests)), triaxial tests, and shear box tests were performed on these samples. In all shear tests, the shear strength measured was found to increase with the inclination &amp / #948 / of the shear plane to the bedding planes. Thus, cylwests (&amp / #948 / = 60o) iii yielded higher values of internal friction &amp / #966 / by about 3.6o than priswests (&amp / #948 / = 30o) under normal stresses between 17 kPa and 59 kPa. Values of &amp / #966 / measured in cylwests were about 1.08 times those measured in triaxial tests (&amp / #948 / &amp / #8776 / 65o), a figure close to the corresponding ratio of 1.13 found by past researchers between actual plane strain and triaxial test results. There was some indication that the difference between cylwest and triaxial test results increased with the &amp / #966 / cv value of the samples. With the smaller &amp / #948 / values (30o and 40o), priswests yielded nearly the same &amp / #966 / values as those obtained in triaxial tests under normal stresses between 20 kPa and 356 kPa. Shear box tests (&amp / #948 / =0o) yielded lower values of &amp / #966 / than cylwests (by about 7.9o), priswests (by about 4.4o), and triaxial tests (by about 4.2o) under normal stresses between 17 kPa and 48 kPa. It was shown that the shear strength measured in shear box tests showed an increase when &amp / #948 / was increased from 30o to 60o / this increase (about 4.2o) was of the order of the difference (about 3.6o) between priswest (&amp / #948 / = 30o) and cylwest (&amp / #948 / = 60o) results mentioned earlier. Shear box specimens with &amp / #948 / = 60o, prepared from the same batch of any sample as the corresponding cylwests, yielded &amp / #966 / values very close to those obtained in cylwests.