No Arabic abstract
We use Topological Data Analysis to study the post buckling behavior of laboratory scale cylindrical silos under gravity driven granular discharges. Thin walled silos buckle during the discharge if the initial height of the granular column is large enough. The deformation of the silo is reversible as long as the filling height does not exceed a critical value, $L_c$. Beyond this threshold the deformation becomes permanent and the silo often collapses. We study the dynamics of reversible and irreversible deformation processes, varying the initial filling height around $L_c$. We find that all reversible processes exhibit striking similarities and they alternate between regimes of slow and fast dynamics. The patterns that occur at the beginning of irreversible deformation processes are topologically very similar to those that arise during reversible processes. However, the dynamics of reversible and irreversible processes is significantly different. In particular, the evolution of irreversible processes is much faster. This allows us to make an early prediction of the collapse of the silo based solely on observations of the deformation patterns.
We investigate, at a laboratory scale, the collapse of cylindrical shells of radius $R$ and thickness $t$ induced by a granular discharge. We measure the critical filling height for which the structure fails upon discharge. We observe that the silos sustain filling heights significantly above an estimation obtained by coupling standard shell-buckling and granular stress distribution theories. Two effects contribute to stabilize the structure: (i) below the critical filling height, a dynamical stabilization due to granular wall friction prevents the localized shell-buckling modes to grow irreversibly; (ii) above the critical filling height, collapse occurs before the downward sliding motion of the whole granular column sets in, such that only a partial friction mobilization is at play. However, we notice also that the critical filling height is reduced as the grain size, $d$, increases. The importance of grain size contribution is controlled by the ratio $d/sqrt{R t}$. We rationalize these antagonist effects with a novel fluid/structure theory both accounting for the actual status of granular friction at the wall and the inherent shell imperfections mediated by the grains. This theory yields new scaling predictions which are compared with the experimental results.
We study the flow of elongated grains (wooden pegs of length $L$=20 mm with circular cross section of diameter $d_c$=6 and 8 mm) from a silo with a rotating bottom and a circular orifice of diameter $D$. In the small orifice range ($D/d<5$) clogs are mostly broken by the rotating base, and the flow is intermittent with avalanches and temporary clogs. Here $dequiv(frac{3}{2}d_c^2L)^{1/3}$ is the effective grain diameter. Unlike for spherical grains, for rods the flow rate $W$ clearly deviates from the power law dependence $Wpropto (D-kd)^{2.5}$ at lower orifice sizes in the intermittent regime, where $W$ is measured in between temporary clogs only. Instead, below about $D/d<3$ an exponential dependence $Wpropto e^{kappa D}$ is detected. Here $k$ and $kappa$ are constants of order unity. Even more importantly, rotating the silo base leads to a strong -- more than 50% -- decrease of the flow rate, which otherwise does not depend significantly on the value of $omega$ in the continuous flow regime. In the intermittent regime, $W(omega)$ appears to follow a non-monotonic trend, although with considerable noise. A simple picture, in terms of the switching from funnel flow to mass flow and the alignment of the pegs due to rotation, is proposed to explain the observed difference between spherical and elongated grains. We also observe shear induced orientational ordering of the pegs at the bottom such that their long axes in average are oriented at a small angle $langlethetarangle approx 15^circ$ to the motion of the bottom.
Granular flow in a silo demonstrates multiple nonlocal rheological phenomena due to the finite size of grains. We solve the Nonlocal Granular Fluidity (NGF) continuum model in quasi-2D silo geometries and evaluate its ability to predict these nonlocal effects, including flow spreading and, importantly, clogging (arrest) when the opening is small enough. The model is augmented to include a free-separation criterion and is implemented numerically with an extension of the trans-phase granular flow solver described in arXiv:1411.5447, to produce full-field solutions. The implementation is validated against analytical results of the model in the inclined chute geometry, such as the solution for the $H_{mathrm{stop}}$ curve for size-dependent flow arrest, and the velocity profile as a function of layer height. We then implement the model in the silo geometry and vary the apparent grain size. The model predicts a jamming criterion when the opening competes with the scale of the mean grain size, which agrees with previous experimental studies, marking the first time to our knowledge that silo jamming has been achieved with a continuum model. For larger openings, the flow within the silo obtains a diffusive characteristic whose spread depends on the models nonlocal amplitude and the mean grain size. The numerical tests are controlled for grid effects and a comparison study of coarse vs refined numerical simulations shows agreement in the pressure field, the shape of the arch in a jammed silo configuration, and the velocity field in a flowing configuration.
An experimental apparatus measuring free-surface deformation of a centrifuged granular pile is developed. By horizontally rotating a quasi two-dimensional granular pile whose apex is located at the vertical rotation axis, the resultant force of gravity and centrifuge yields the deformation of the granular pile. In this setup, centrifugal force depends on distance from the rotation axis whilst gravitational force is constant everywhere. Therefore, free-surface deformation by various centrifuge degrees can be systematically examined using this apparatus. In the system, a small unit consisting of a camera and computer is rotated with the granular sample to record the rotation-induced deformation. To evaluate the validity of the system, deformation of a rotated water surface is first measured and analyzed. The obtained data are properly explained by the theoretical parabolas without any fitting parameter. Next, we measure the deformation of non-cohesive and cohesive granular piles using the developed apparatus. Both granular samples show the significant deformation of granular pile and finally develop steep granular slopes on the side walls. However, details of the deformation processes depend on the cohesion strength. To quantitatively characterize the difference, the effective strength by cohesion and granular local-slope variations are analyzed based on the experimental results.
We study experimentally gravity-driven granular discharges of laboratory scale silos, during the initial instants of the discharge. We investigate deformable wall silos around their critical collapse height, as well as rigid wall silos. We propose a criterion to determine a maximum time for the onset of the collapse and find that the onset of collapse always occurs before the grains adjacent to the wall are sliding down. We conclude that the evolution of the static friction toward a state of maximum mobilization plays a crucial role in the collapse of the silo.