Scale dependence of direct shear tests

Direct shear test has been widely used to measure the shear strength of soils and other particulate materials in industry because of its simplicity. However, the results can be dependent on the specimen size. The ASTM (American Society for Testing and Materials) publications suggest that for testing soils the shear box should be at least ten times the diameter of the largest particle and the height of the box should be no more than half of its diameter. These guidelines are empirically based. A series of two-dimensional numerical direct shear tests are performed to investigate this scaling effect. By analyzing the bulk friction, particle translation and rotation, percentage of sliding, average volume (area) and shear strain and the evolution of the shear band, we find that the traditional guidelines for direct shear tests are questionable. Scaling dependency of bulk friction on the property of granular materials is clearly present. Our current analysis points out that the scaling effects can vary significantly depending on the particle properties other than their sizes. Of all the parameters we observed, particle rotation appears to have a decisive correlation with the bulk friction. Formation of a shear band is universal. As the shearing progresses, particle rotation begins to concentrate near the shear plane. By defining the width of a shear band as the standard deviation of the distribution of translational gradient or the standard deviation of the distribution of particle rotation, quantitative evolutions of shear band are presented. Both measures of the shear band width dropped rapidly during pre-failure stage. After peak stress both measures begin to approach steady state as the bulk friction stabilizes to the residual stage. These observations suggest that structure formation inside the shear band controls the scaling effect.