Taira YUMORI Development of liquefaction testing method under controlled principal stress directions of cyclic shearing Hirofumi Toyota Previous studies have indicated that particle orientation and major principal stress direction affect the liquefaction characteristics of soil. However, it is difficult to fully evaluate the relationship between particle orientation and principal stress direction in the conventional triaxial test because the principal stress direction is limited to control either vertical or horizontal. Moreover, the intermediate principal stress b, which can be controlled in the triaxial test, is limited to 0 (compression) or 1 (extension). Therefore, it is not possible to evaluate the effects of particle orientation and principal stress direction on liquefaction characteristics under 3D stress conditions. In previous studies, liquefaction tests using triaxial tests and shaking table tests have been conducted on sand samples with particle orientation. Those two testing results presented a discrepancy, in which liquefaction strength trend induced by particle orientation is opposite each other. This suggests that differences in loading method, i.e., principal stress direction and intermediate principal stress state, may affect liquefaction behavior. Therefore, it is necessary to evaluate liquefaction characteristics under conditions to evaluate the difference between particle orientation and major principal stress direction. In this study, a hollow torsional shear apparatus was used to develop monotonic loading tests and liquefaction tests in which both the principal stress direction and intermediate principal stress coefficient b could be controlled. Toyoura sand specimens prepared with the same depositional angle were isotropically consolidated. Then, the specimens were sheared under a constant mean principal stress condition with the angle ƒ¿, which is the maximum principal stress direction from the vertical. The ƒ¿was varied in the range of 0‹ to 90‹ (ƒ¿ = 0‹ indicates a state in which the major principal stress acts vertically, and ƒ¿ = 90‹ indicates a state acting horizontally). Furthermore, the intermediate principal stress coefficients b was also varied to comprehensively evaluate the effects of 3D stress conditions. As a result, under monotonic loading, clear differences were observed in stress path and stress?strain relation: When ƒ¿ or b was greater, pore water pressure generated largely and the incarnation of the failure line became gentle. However, in the liquefaction test, the influence of ƒ¿ was reduced because the principal stress direction reversed by 90‹ between forward and reverse rotations during the cyclic loading. @