This study explores the features of motions at contact scale, associated with the evolution of fabric, meso-scale structures and deformation mechanisms in 2D assemblies subjected to shearing. A tailor-made direct shear box and a specially designed biaxial cell with a flexible boundary have been set-up. The Particle Image Velocimetry (PIV) technique and close-range photogrammetry are used together to help with the investigation. In the direct shearing, the test sample is made of randomly packed wooden rods of three different diameters. In the biaxial testing, the elliptical rods and the object dots on the rods, produced by the 3D printer, are used as the test samples to improve the resolution of the subsequent image analyses. The 3D printing technique is also applied to ease man...[ Read more ]

This study explores the features of motions at contact scale, associated with the evolution of fabric, meso-scale structures and deformation mechanisms in 2D assemblies subjected to shearing. A tailor-made direct shear box and a specially designed biaxial cell with a flexible boundary have been set-up. The Particle Image Velocimetry (PIV) technique and close-range photogrammetry are used together to help with the investigation. In the direct shearing, the test sample is made of randomly packed wooden rods of three different diameters. In the biaxial testing, the elliptical rods and the object dots on the rods, produced by the 3D printer, are used as the test samples to improve the resolution of the subsequent image analyses. The 3D printing technique is also applied to ease manufacture of the biaxial device.

In the direct shearing, it is found that both the rolling and sliding components are concurrent at the same contacts. Both components are simultaneously increased and gradually concentrated on some particular contacts during shearing. This suggests bimodal behavior of the movements at the contacts, i.e., a strong motion contact versus a weak motion contact. Strong motion contacts are essential to the volumetric responses; contact sliding produces volumetric contraction whereas contact rolling leads to volumetric dilation. As the sample volume continues to dilate, the averaged cumulative rolling distance gradually prevails over the cumulative sliding distance and therefore the difference between these two distances is steadily increased. Such a difference, however, is reduced as the applied vertical stress increases because a higher confining pressure hinders the development of rolling-induced dilation by simultaneously promoting the counter effect from contact sliding between the roll-over particle pair.

The biaxial testing is employed in the study of fabric and meso-scale structures. The Set Voronoi diagram is used to partition the assembly. The particle cell is established based on the Set Voronoi cell to calculate the local strain of which the effect of particle rotation is considered. A good agreement between the measured and calculated volumetric responses is found, thereby supporting the effectiveness of the local strain calculation used in this study. In derivation of the void-based fabric tensor, an image-based method is used to extract the void space, which is then represented by the best-fit ellipse. This fitting ellipse is the basis to calculate the void-based fabric tensor of which the configuration of the ellipse is incorporated. A strong linear relationship among the three fabric anisotropy intensity factors, i.e., the particle orientation-based, contact normal-based and void-based can also be found in samples. Based on the proper orthogonal decomposition (POD) on the nonaffine motion fields, it is found that the evolution of coherent vortex structures correlate well with the formation process of shear band.

The biaxial testing is also used in the study of deformation mechanisms. The motion of a particle pair is divided into three types, i.e., the contact deformation, the Type 4 rolling, and the rigid motion. Each of the motion has different contributions to the sample deformation at different strain levels. The normal component of contact deformation (i.e., rotational sliding) dominates volumetric contraction at small strains and dilation at larger strains. The local volumetric strains arising from this type of motion mainly take place at the interfaces between moving particle clusters and vortices, and at the boundaries of micro bands or shear bands. Since the elliptical rods are used in the experiment, the rigid body rotation is hindered. Therefore, the contribution of the rigid body rotation changes from slight dilation to contraction when the particle shape becomes more elongated. For the sample distortion, it is due mainly to the rigid body rotation and the normal component of contact deformation makes a second contribution. The tangential component of contact deformation (i.e., translational sliding) and Type 4 rolling relatively have small contribution to both volumetric strain and sample distortion.