betaFeedback

WorksPhysics and Astronomy

Internal Waves Control Bulk Flow in Silos

Abstract

We experimentally measure paticle acceleration within the bulk during the discharge of a granular silo. We highlight the existence of a deceleration wave emerging at the outlet level near the dead zone and propagates toward the top of the medium. The wave emission frequency is extracted from spatiotemporal diagrams of the Eulerian instantaneous acceleration profiles. Surprisingly, we find that this frequency decreases with the cohesion of the medium and is independent of the outlet size.

Granular media frequently show complex behavior in a variety of geophysical phenomena and industrial applications, such as additive manufacturing or food processing [1, 2, 3]. Among the various applications of granular materials, silos are one of the most commonly used types of equipment in the industry for storage and handling [4, 5, 6, 7, 8, 9, 10, 11]. Amidst the complexity of hopper flows, which have been extensively studied and characterized through the identification of semi-empirical laws governing various aspects, such as the influence of aperture size on flow rate [4, 5, 10] and the morphology of velocity profiles both at the outlet [12, 13] and in the bulk [7, 11], the occurrence of “silo quaking” emerges as a recurrent challenge in the realm of silo discharge dynamics. Gravity-driven flows within silos typically manifest as a cyclic or pulsating motion, influenced by fluctuations in material density during flow and varying degrees of internal friction mobilization and boundary wall friction [14, 15, 16]. This pulsating flow phenomenon becomes notably pronounced at lower flow rates, exhibiting periods ranging from mere seconds to extended durations, potentially extending to minutes [17, 18, 19]. The repercussions of such pulsating loads vary, ranging from minor nuisances like shock wave transmission through the ground to disturbances in neighboring areas, to the more severe structural fatigue failures induced when the flow pulses excite the natural frequencies of the silo and its supporting structure [20, 21]. A related issue to the silo quaking problem is the phenomenon known as “silo music” and “silo honking,” as reported by Tejchman and Gudehus [22]. Silo honking, specifically, is observed in tall aluminum silos storing plastic powders, where higher-frequency components of flow pulsations generate loud, periodic “fog-horn” type sounds, causing notable disturbances [23].

Through an experiment conducted in a 2D silo, we demonstrate the existence of internal waves and characterize them. Specifically, we examine the frequency of these waves and show that it is independent of the aperture size, contrary to the scaling laws describing the flow rate. Additionally, we find that the frequency of these waves is strongly influenced by the cohesion of the medium. We interpret these results in light of the formation of free-fall arch in the lower region of the silo, at the interface between the flowing and the dead zones.

The experimental setup is a quasi-2D silo (shown in Figure 1 (b)) identical to that used by Pascot et al. [24, 25] in their study of the influence of vibrations on silo discharge. It consists of two conductor polycarbonate plates which prevent electrostatic effects. The dimensions of the cavity are H=300H = 300 mm (height), L=100L = 100 mm (width) and W=1.25W = 1.25 mm (depth) with an adjustable aperture size DD set here from 6 mm to 24 mm. The narrow depth of the silo allows for a monolayer of amorphous 2D stack of spherical particles. When the silo is discharging, grains fall directly into a collecting container.

Flow near the silo outlet and schematic of the experimental setup

Figure 1. (a) Flow near the silo outlet with the corresponding velocity profile (shown in red). Black dashed lines indicate the boundaries of the dead zones, defined by v/gd<0.05v/\sqrt{gd} < 0.05. (b) Schematic of the experimental setup.

The flow is recorded by a high-speed camera with a frame rate of 2000 Hz and exposure time of 498 μs. Pictures (1024×1024 pixels) show a 10×1010 \times10 cm2^2 area at the bottom of the quasi-2D silo (see Figure 1 (a)). Grains are uniformly illuminated and individually identified using a particle tracking algorithm. The granular media used is composed of bidisperse steel beads of diameter 1.0 mm and 1.2 mm (mean diameter d=1.1d = 1.1 mm ±0.1\pm0.1 mm). Coils in a Helmholtz configuration are used to apply a uniform magnetic field across the silo opening. When placed in a magnetic field, beads acquire magnetic dipolar moments and behave like small magnets. They attract each other in the vertical direction and repel each

Typeset by REVTeX other in the horizontal direction. The induced cohesion is controlled by adjusting the current II passing through the coils (up to 4 A here).

To quantify interparticle cohesion, we define a magnetic Bond number as the ratio of the magnetic force to the gravitational weight of a single grain, following the standard approach used in magnetic cohesion–dominated systems [26, 27, 28]. By measuring the angle of repose of the dead zones formed after discharge, which ranges from 28∘28^\circ to 55∘55^\circ, we estimate Bond numbers between 0 and 51.1, consistent with previously reported results [28].

In this work, we operate in a regime where the outlet size is sufficiently large to prevent complete flow arrest due to the formation of stable arches. Unlike previous studies [18, 29, 30, 31] focusing on the jamming events occurring at the silo exit, our analysis examines flow dynamics within the bulk of the granular material. In this region, we observe velocity profiles exhibiting the characteristic shape shown in Figure 1 (b), with a central, channelized flowing zone bordered by two quasi-static dead zones.

Figure 2 (a,Figure 2b) show the evolution of the flow rate QQ and the volume fraction ϕc\phi_c at the center of the outlet as a function of the relative opening D/dD/d. In the absence of cohesion, these quantities follow the classical laws obtained by Beverloo and Benyamine [4, 5, 32], respectively. QQ decreases slightly with $mathBoBo, as does ϕc\phi_c, which reaches an asymptotic limit at large openings, diminishing further as the Bond number increases. Cohesion has no significant effect on the velocity and compaction profiles at the outlet, which remain parabolic and self-similar [12] across all Bond numbers (see Figure 2 (a)).

Flow rates, normalized density, velocity, and density profiles

Figure 2. (a) Flow rates with Beverloo law for C=0.65C = 0.65, k=1.5k = 1.5 and ϕb=0.58\phi_b = 0.58 and (b) normalized density ϕc(x)/ϕb\phi_c(x)/\phi_b for all values of aperture size D/dD/d and Bond number BoBo (α=0.66\alpha= 0.66, β=0.11\beta= 0.11 and ϕb=0.58\phi_b = 0.58). Normalized profiles of (c) velocity and (d) density at the outlet level normalized by their fitted center value vcv_c and ϕc\phi_c.

The previous results are derived from the mean flow velocity during silo discharge. Here, using the high spatiotemporal resolution of our particle tracking, which provides low-noise second derivatives of particle positions, we analyze the instantaneous acceleration field to characterize flow fluctuations.

Figure 3 (a) shows the spatio-temporal diagram of the vertical acceleration ay(x=0,y,t)a_y(x = 0, y, t) of the grains located at the center of the silo during discharge in the absence of cohesion. Clear intermittency is observed, highlighting the existence of waves propagating from the opening toward the top of the silo. Figure 3 (c) illustrates the influence of cohesion on these waves, showing that the duration of acceleration and deceleration phases increases with cohesion.

Figure 3. Spatiotemporal diagrams and corresponding power spectrum of the instantaneous acceleration ay(x=0)a_y(x = 0) for Bo=0Bo = 0 (a,b) and Bo=51.1Bo = 51.1 (c,d). D=24D = 24 mm in both cases.

The waves are generated above the outlet, at heights ranging from 5 to 20 dd. These acceleration waves neither block the silo nor significantly alter the mean flow rate. The flow below the region where these waves are generated exhibits particle accelerations nearly equal to the gravitational acceleration. To determine the characteristic wave frequency, we compute the power spectrum of each line of the spatio-temporal diagrams ay(x=0,y,t)a_y(x = 0, y, t), using Welch’s method (see Figure 3 (b,Figure 3d)). The waves exhibit a clear periodicity with a well-defined frequency, about one-third of that in the non-cohesive case (∼8\sim8 Hz vs. ∼25\sim25 Hz).

Examples of instantaneous snapshots (Figure 4 (b-Figure 4f)) show that these waves originate at the edges of the dead zones, on either side. These phenomena resemble the “silo honking” or “silo music” effect [22], where sudden variations in grain flow produce waves and intermittency. Here, however, the effect does not stem from particle synchronization or wall-induced resonance, but from the periodic formation of dynamic force chains with lifetimes Fig. 5 (a) presents spatial maps of the normalized velocity fluctuations v′(x,y)/gdv'(x,y)/\sqrt{gd} within the silo near the aperture. These maps clearly reveal transient dynamic arches, which are located progressively higher in the silo as the aperture becomes larger. Remarkably, the zone of intensified fluctuations forming each arch aligns with a contour of constant acceleration (shown in black on Figure 5 (a)), demonstrating the direct relationship between local velocity fluctuations and the acceleration field associated with these temporary structures. Fig. 5 (b,c) show, respectively, the evolution of the arch height h∩/dh_{\cap}/d and of the associated iso-acceleration value a∩/ga_{\cap}/g as a function of the aperture size. These results clearly indicate that the arches form at increasingly higher positions and become progressively less stable as the aperture increases.

Spatiotemporal diagram and instantaneous acceleration profiles

Figure 4. (a) Example of a spatiotemporal diagram with (b-f) corresponding instantaneous acceleration profiles ay(x,y,t)/ga_y(x,y,t)/g for D=24D=24 mm and Bo=0Bo=0.

Profiles of normalized velocity fluctuations and transient arch properties

Figure 5. (a) Profiles of the normalized velocity fluctuations, v′(x,y)/gdv'(x,y)/\sqrt{gd}, for Bo=0Bo=0 and different aperture ratios D/dD/d. (b) Dependence of the transient arch height, h∩/dh_{\cap}/d, on the silo opening. (c) Dependence of the iso-acceleration threshold of transient arches, a∩/ga_{\cap}/g, on the silo opening. The star symbol indicates the limiting case for which the iso-acceleration equals gg. Solid lines in panels (b) and (c) are guides to the eye.

These arches act as the sources of the emitted waves. We observe that even though cohesion does not have a significant influence on the flow rate, it greatly alters the emission frequency of these waves. This effect is highlighted in Figure 6, which shows the evolution of the relative frequency f/f0f/f_0 as a function of the Bond number, where f0f_0 is the frequency of the waves at Bo=0Bo=0. f/f0f/f_0 decreases as cohesion increases, showing that deceleration waves are affected by cohesion in the bulk flow. Surprisingly, we observe that ff and f0f_0 do not depend on the aperture size DD suggesting that these waves are controlled by the grain size scale. By supposing that the magnetic field induces the formation of oriented aggregates along the vertical direction with a characteristic length ξ=d(1+Bo/Boc)\xi=d(1+Bo/Bo_c) associated with a critical Bond number BocBo_c, the characteristic frequency related to the flow scales as f∼g/ξf\sim\sqrt{g/\xi}. This implies that:

Evolution of the relative frequency of internal acceleration waves as a function of the Bond number

Figure 6. Evolution of the relative frequency f/f0f/f_0 of internal acceleration waves as a function of the Bond number, BoBo, for various silo openings. Black dashed line is given by Eq. (1) for Boc=7Bo_{\mathrm{c}}=7. Insert: f0f_0 as a function of aperture size, DD. The red dashed line corresponds to the mean value ⟨f0⟩=25.8\langle f_0\rangle=25.8 Hz.

f/f0∼(1+Bo/Boc)−0.5(1)f/f_0 \sim(1+Bo/Bo_c)^{-0.5} \tag*{(1)}

and that the blocking frequency is determined by the free falling time under the effect of gravity of the clusters over a spatial scale equal to their diameter ξ\xi. This model is shown in Figure 6 and describes the decay of the frequency shorter than the flow’s characteristic timescale.

relatively well. It is important to note that the model presented here is a purely geometrical model where the details of the dynamics are embedded in f0f_0, the frequency of the waves in the absence of cohesion found to be 25.8 Hz in our experiments.

This frequency can be compared to the sound emission frequency observed in the phenomenon of booming sand dunes [33, 34], which has been experimentally found to be on the order of αg/d\alpha\sqrt{g/d} with α=0.4\alpha=0.4 [35] over a wide range of particle sizes and flow angles, giving f0=37.7f_0=37.7 Hz in our experiments. This value is very close to the one we obtained, and the difference between the two values arises from the fact that the prefactor $\alpha depends both on the local angle of the slope and the dissipation coefficient during collisions [36] which are difficult to measure experimentally. This result suggests that the frequency of these waves in silos is associated with the velocity gradient at the surface of the dead zones, where the granular medium periodically dilates and compacts. This implies that arches form on a timescale corresponding to the time it takes for a grain to move by one diameter. The effect of cohesion can then be taken into account by assuming the formation of particle clusters whose size increases linearly with the Bond number.

Our results highlight the existence of traffic waves during silo discharge, even though the outlet flow rate remains steady. These waves are generated near the outlet by dynamic arches that locally block the flow over a characteristic timescale roughly equal to the time it takes for a grain to move by one diameter. Although this timescale is short, it scales with the grain’s free-fall time downstream of the arch, ensuring that the macroscopic flow rate remains constant. These waves propagate along the dead zones up to the top of the packing through traffic wave dynamics similar to that observed in booming sand dunes. Here, we demonstrate for the first time that the effect of cohesion can be incorporated into this mechanism by assuming the formation of clusters whose size increases with the Bond number. Our results suggest the possibility of indirectly measuring the cohesion of a granular medium by measuring the frequency of these waves. It remains an open question whether such waves occur in real cohesive powders, and how their propagation over large heights, amplification, and interaction with silo walls affect the homogeneity of silo discharge.

This work was financially and scientifically supported by the Lorraine Université d’Excellence (LUE) PIA project and the Institut Universitaire de France, whose contributions were essential to this investigation.

References

  1. [1]B. Andreotti, Y. Forterre, and O. Pouliquen, Granular Media: Between Fluid and Solid (Cambridge University Press, 2013).
  2. [2]D. Barletta, G. Donsì, G. Ferrari, M. Poletto, and P. Russo, Solid flow rate prediction in silo discharge of aerated cohesive powders, AIChE Journal 53, 2240 (2007).DOI
  3. [3]A. Cannavacciuolo, D. Barletta, G. Donsì, G. Ferrari, and M. Poletto, Arch-free flow in aerated silo discharge of cohesive powders, Powder Technology 191, 272 (2009).DOI
  4. [4]G. Hagen, Über den Druck und die Bewegung des trocknen Sandes, Bericht über die zur Bekanntmachung geeigneten Verhandlungen der Königlich Preussischen Akademie der Wissenschaften zu Berlin , 35 (1852).
  5. [5]W. Beverloo, H. Leniger, and J. Van de Velde, The flow of granular solids through orifices, Chemical Engineering Science 15, 260 (1961).DOI
  6. [6]J. Litwiniszyn, An application of the random walk argument to the mechanics of granular media, Rheology and Soil Mechanics: Proceedings of the Grenoble Symposium, April 1–8, 1964 , 82 (1964).
  7. [7]W. Mullins, Experimental evidence for the stochastic theory of particle flow under gravity, Powder Technology 9, 29 (1974).DOI
  8. [8]R. Nedderman and U. Tüzün, A kinematic model for the flow of granular materials, Powder Technology 22, 243 (1979).
  9. [9]A. Medina, J. Cordova, E. Luna, and C. Trevino, Velocity field measurements in granular gravity flow in a near 2d silo, Physics Letters A 250, 111 (1998).DOI
  10. [10]C. Mankoc, A. Janda, R. Arevalo, J. Pastor, I. Zuriguel, A. Garcimartín, and D. Maza, The flow rate of granular materials through an orifice, Granular Matter 9, 407 (2007).
  11. [11]M. Bazant, A theory of cooperative diffusion in dense granular flows, arXiv preprint cond-mat/0307379 (2003).DOI
  12. [12]A. Janda, I. Zuriguel, and D. Maza, Flow rate of particles through apertures obtained from self-similar density and velocity profiles, Physical Review Letters 108, 248001 (2012).pubmed.ncbi.nlm.nih.gov/23004334
  13. [13]M. Benyamine, M. Djermane, B. Dalloz-Dubrujeaud, and P. Aussillous, Discharge flow of a bidisperse granular media from a silo, Phys. Rev. E 90, 032201 (2014).pubmed.ncbi.nlm.nih.gov/25314431
  14. [14]M. Sperl, Experiments on corn pressure in silo cells–translation and comment of janssen’s paper from 1895, Granular Matter 8, 59 (2006).DOI
  15. [15]Z. Xu and P. Liang, Modified lateral pressure formula of shallow and circular silo considering the elasticities of silo wall and storage materials, Scientific Reports 12, 7069 (2022).
  16. [16]C. Perge, M. Aguirre, P. Gago, L. Pugnaloni, D. Le Tourneau, and J.-C. Géminard, Evolution of pressure profiles during the discharge of a silo, Physical Review E 85, 021303 (2012).pubmed.ncbi.nlm.nih.gov/22463198
  17. [17]I. Zuriguel, A. Garcimartín, D. Maza, L. Pugnaloni, and J. Pastor, Jamming during the discharge of granular matter from a silo, Phys. Rev. E 71, 051303 (2005).pubmed.ncbi.nlm.nih.gov/16089524
  18. [18]I. Zuriguel, D. Parisi, R. Hidalgo, C. Lozano, A. Janda, P. Gago, J. Peralta, L. Ferrer, L. Pugnaloni, E. Clément, et al., Clogging transition of many-particle systems flowing through bottlenecks, Scientific reports 4, 1 (2014).DOI
  19. [19]I. Zuriguel, D. Maza, A. Janda, R. Hidalgo, and A. Garci-martín, Velocity fluctuations inside two and three dimensional silos, Granular Matter 21, 1 (2019).DOI
  20. [20]J. Carson, Silo failures: Case histories and lessons learned, Handbook of Powder Technology 10, 153 (2001).DOI
  21. [21]K. Wilde, M. Rucka, and J. Tejchman, Silo music mechanism of dynamic flow and structure interaction, Powder Technology 186, 113 (2008).DOI
  22. [22]J. Tejchman and G. Gudehus, Silo-music and silo-quake experiments and a numerical cosserat approach, Powder Technology 76, 201 (1993).DOI
  23. [23]B. Muite, S. Quinn, S. Sundaresan, and K. Rao, Silo music and silo quake: granular flow-induced vibration, Powder Technology 145, 190 (2004).
  24. [24]A. Pascot, N. Gaudel, S. Antonyuk, J. Bianchin, and S. Kiesgen de Richter, Influence of mechanical vibrations on quasi-2d silo discharge of spherical particles, Chemical Engineering Science 224, 115749 (2020).DOI
  25. [25]A. Pascot, J.-Y. Morel, S. Antonyuk, M. Jenny, Y. Cheny, and S. Kiesgen de Richter, Discharge of vibrated granular silo: A grain scale approach, Powder Technology 397, 116998 (2022).DOI
  26. [26]N. Vandewalle and G. Lumay, Flow properties and heap shape of magnetic powders, AIP Conference Proceedings 1145, 135 (2009).DOI
  27. [27]L. Thorens, M. Viallet, K. J. Måløy, M. Bourgoin, and S. Santucci, Discharge of a 2d magnetic silo, EPJ Web of Conferences 249, 03017 (2021).
  28. [28]A. J. Forsyth, S. R. Hutton, M. J. Rhodes, and C. F. Osborne, Effect of applied interparticle force on the static and dynamic angles of repose of spherical granular material, Physical Review E 63, 031302 (2001).pubmed.ncbi.nlm.nih.gov/11308644
  29. [29]K. To, P.-Y. Lai, and H. Pak, Jamming of granular flow in a two-dimensional hopper, Phys. Rev. Lett. 86, 71 (2001).pubmed.ncbi.nlm.nih.gov/11136096
  30. [30]D. Gella, D. Maza, I. Zuriguel, A. Ashour, R. Arévalo, and R. Stannarius, Linking bottleneck clogging with flow kinematics in granular materials: The role of silo width, Phys. Rev. Fluids 2, 084304 (2017).DOI
  31. [31]A. Nicolas, A. Garcimartín, and I. Zuriguel, Trap model for clogging and unclogging in granular hopper flows, Phys. Rev. Lett. 120, 198002 (2018).DOI
  32. [32]M. Benyamine, P. Aussillous, and B. Dalloz-Dubrujeaud, Discharge flow of a granular media from a silo: effect of the packing fraction and of the hopper angle, EPJ Web of Conferences 140, 03043 (2017).
  33. [33]B. Andreotti, The song of dunes as a wave-particle mode locking, Physical Review Letters 93, 238001 (2004).pubmed.ncbi.nlm.nih.gov/15601204
  34. [34]N. Vriend, M. Hunt, R. Clayton, C. Brennen, K. Brantley, and A. Ruiz-Angulo, Solving the mystery of booming sand dunes, Geophysical Research Letters 34 (2007).
  35. [35]S. Dagois-Bohy, S. Courrech du Pont, and S. Douady, Singing-sand avalanches without dunes, Geophysical research letters 39 (2012).
  36. [36]L. Quartier, B. Andreotti, S. Douady, and A. Daerr, Dynamics of a grain on a sandpile model, Physical Review E 62, 8299 (2000).pubmed.ncbi.nlm.nih.gov/11138128

Paper details

vlnc:W_5zwqw57c
Version
v1 · latest
Published
Fresh — added to the Hub within the last 30 days
Added
License
CC BY 4.0
Usage

views · 0

downloads · 0

A view counts once per browser per day, a download each time the PDF is requested.

  • Physics and Astronomy

Not recorded for this paper

View statement
Looking for related papers…
Contents