Get 20M+ Full-Text Papers For Less Than $1.50/day. Start a 14-Day Trial for You or Your Team.

Learn More →

Kinetic theory based segregation model for density-bidisperse dense granular flows

Kinetic theory based segregation model for density-bidisperse dense granular flows Kinetic theory based segregation model for density-bidisperse dense granular ows 1 2 1, 2, 3 1, 2, 3, Yifei Duan, Paul B. Umbanhowar, Julio M. Ottino, and Richard M. Lueptow Department of Chemical and Biological Engineering, Northwestern University, Evanston, Illinois 60208, USA, Department of Mechanical Engineering, Northwestern University, Evanston, Illinois 60208, USA Northwestern Institute on Complex Systems (NICO), Northwestern University, Evanston, Illinois 60208, USA, (Dated: June 11, 2019) Individual constituent balance equations are often used to derive expressions for species speci c segregation velocities in ows of dense granular mixtures. We propose an expression for the in- terspecies momentum exchange in density bidisperse granular ows as an extension of ideas from kinetic theory, expanding its range from short-duration binary collisions to the multiple enduring contacts characteristic of dense shear ows, and incorporating the e ects of particle friction, con- centration ratio, and local ow conditions. The segregation velocity derived from the momentum balance equation with the new interspecies drag model matches results from DEM simulations and accurately predicts the downward and upward segregation velocities of heavy and light particles through the depth of the owing layer for di erent density ratios and constituent concentrations for both con ned shear and free surface heap ows. I. INTRODUCTION When sheared, dense granular materials tend to segregate by particle size, density, shape or friction coecient, which can be problematic in many industries due to its impact on product quality and uniformity [1{8]. As a result, the segregation behavior of sheared granular mixtures is of interest in a wide variety of elds from both fundamental and applied standpoints. We focus on bidisperse granular mixtures of particles having the same size but di erent densities such that segre- gation is driven by a \buoyant force" mechanism [9, 10]. ln this paper, we propose a model for the interspecies drag between heavy and light particles in density bidisperse granular ows and use the model to derive an expression for the segregation velocity of each species by considering the balance between the net buoyant force and the interspecies drag force. To model the segregation process, we assume that two di erent particle species can be treated as interpenetrable continua such that each constituent has its own partial pressure, consistent with previous studies [1, 3, 4]. The imbalance between a species' partial pressure and its body force results in segregation of the two species. A linear relation between the drag force on a particle and its segregation velocity has been used in several gravity-driven segregation models for size or density segregation [1, 4, 11{14] based on either the similarity of kinetic sieving in granular segregation with uid percolation through a porous material or an analogy to the drag force on a sphere settling in a uid. The linear drag model has been combined with the constituent mass and momentum balances to derive a general multicomponent theory for segregation [15]. However, Weinhart et al. [16] demonstrated that the coecient tted from the linear drag model varies with time, indicating that a simple linear drag law fails to describe the segregation behavior. Recent DEM simulations also suggest that other factors, such as the local pressure [17, 18] and the species concentration ratio [19, 20], a ect the segregation velocity, which again raises questions concerning the validity of a linear drag model. As an alternative to the linear drag model, it is natural to consider the Kinetic Theory of Granular Flow (KTGF) to model the momentum transfer between two segregating species. The KTGF provides a drag model connecting stresses and velocities in a granular mixture while accounting for complex particle interactions [21{27], thereby yielding a drag model that takes particle density and diameter into consideration at the particle level. However, there are two major problems with using the KTGF approach to model dense ows. First, the stress generation mechanism in the KTGF is based on short-duration, binary collisions typical of dilute particle ows as opposed to multiple, enduring contacts typical of the dense granular ows considered here. Second, the KTGF approach does not include physics known to be critical to segregation in dense granular ows such as the dependence of segregation on the local pressure [17]. Nevertheless, a potential path toward a robust drag model for dense ows lies in modifying the KTGF drag model based on empirical results for density segregation in dense granular ows. In this paper, we propose an expression for the density segregation velocity based on mixture theory [1] with a drag model based on KTGF but extended to the dense regime by introducing appropriate coecients. The proposed drag Corresponding author: r-lueptow@northwestern.edu arXiv:1906.03378v1 [cond-mat.soft] 8 Jun 2019 2 model is described in Section II, and an expression for the segregation velocity in density bidisperse granular ow is derived by combining the drag and the equilibrium momentum balance equation. The e ects of the local inertial number, particle friction, and relative species concentration proposed in the model are tested separately by DEM simulations in Section III. In Section IV, results of the segregation model are compared with DEM simulations for ows under more general conditions. Conclusions are given in Section V. II. KTGF BASED DRAG MODEL AND SEGREGATION VELOCITY For dense granular ows, the total solids volume fraction,  =  , where  represents the local volume solid i i fraction of each solid species (here i = h for heavy particles and i = l for light particles), is around 0.6. The local constituent concentration is c =  = and the solid density is  =  c , where  is the particle material i i solid solid i i i density. The gradient of the lithostatic pressure P is @P =   g; (1) solid solid @z where z is the vertical coordinate (assuming negative z is the gravity direction) and g is the acceleration due to gravity. The partial pressure P is determined by partitioning the lithostatic pressure induced by gravity among the two constituents such that P = P . For particles that only di er in density, we assume that the proportion of the hydrostatic load carried by each species equals its concentration c regardless of the density di erence as proposed by Marks et al. [14] and Tunuguntla et al. [4]. The momentum balance for each species can be written as [4] (  uuu ) +rrr (  uuu uuu ) = c rrrP +   ggg + ; (2) i i i i i i i i i i i @t where is the dyadic product, t is time, uuu is the vector velocity, and is the interspecies drag vector. (Eq. (2) is i i similar to Eq. (2.5) in [4], noting that there the intrinsic density  is the bulk density of particles in a non-mixed i i i state, which is equal to   , and that  = c  is equivalent to   used here.) Assuming that the vertical i solid i i acceleration terms are negligible [4, 28], which is reasonable for the relatively slow segregation in a typical granular owing layer, the momentum conservation equation in the z-direction can be simpli ed using Eq. (1) to (  ) g + = 0: (3) solid i i i Equation (3) is analogous to a simple force balance between the the net buoyant force (  ) g and the solid i i interspecies drag force . To determine constituent velocities from Eq. (3), must be expressed as a function of the velocity di erence between the two species, and the KTGF o ers a means to do this. Changes to the KTGF are needed, however, because it was developed to model rapid ow of dilute granular materials based on instantaneous binary collisions [29, 30]. This is not the case in the dense ow regime where enduring contacts dominate and many adjacent particles are part of contact force chains [31]. Correlations in motion and force due to the dense contact network reduce the collisional energy dissipation [32]. Substantial e ort has been expended to extend the applicability of the KTGF to the dense regime, mainly by modifying the expression for the collisional energy dissipation rate to account for the e ects of sustained contacts [33{37]. In fact, an augmented KTGF approach has been proposed for segregation problems, but it is limited to small size and density di erences [38, 39]. In addition, several stand-alone drag models based on the KTGF have been proposed for dilute granular ows [21{27]. Here we follow an approach similar to one of these models [27] by assuming that the long-duration collisional drag follows similar physics to the short-duration collisional drag, but with a much longer contact time. This can be represented by correction coecients modifying the well-known solid-solid drag model proposed by Syamlal [22] that has been successfully used for uidized bed simulations [40{42]. The Syamlal expression for the local drag between two constituents i and j is 3(1 + e) = + g     (d + d ) juuu uuu j(uuu uuu ); (4) i ij i j i j i j i j i j 3 3 2( d +  d ) 2 8 i j i j where g is the radial distribution function for a collision between particle i and j [43],  is the friction coecient, ij and e is the restitution coecient. For density segregation, d = d , and g becomes a constant if we assume i j ij constant total solid volume fraction  . The bulk velocity in the vertical direction of the combined constituents solid 3 is w = c w + c w based on volume conservation. The segregation, or percolation velocity w , for species i is bulk i i j j p;i de ned as the velocity di erence between the constituent velocity and the bulk velocity in the z direction: w = w w = c (w w ): (5) p;i i bulk j i j Using Eq. (5) in Eq. (4) for bidisperse granular mixtures of heavy and light particles having the same diameter d, the interspecies drag force in the segregation (z) direction is jw j h l i p;i = 3(1 + e) 1 + g  w : (6) i ij solid p;i 4 ( +  ) d h l j Based on Eq. (6), we propose a semi-empirical interspecies drag model of the form local concentration ratio KTGF z }| { z }| { jw j 1 h l i p;i h = B()  w : (7) i solid p;i | {z } ( +  ) d I l h j l |{z} enduring contacts local ow Equation (7) has four multiplicative terms. Consider them from left to right. B() replaces the rst term in square brackets in Eq. (6) to account for enduring contacts in dense ows. In dense granular mixtures short-duration collisions are unlikely to play as important of a role as in the dilute ows upon which Eq. (6) for KTGF is based. As a result, we assume B() is independent of e, which is proven in Section III.D. The second term is inherited from the KTGF and re ects the momentum transfer between two particles with di erent mass, taking into consideration density and diameter at the particle level. The third term accounts for the local ow conditions, since Eq. (6) does not explicitly account for the presence of shear. As such, Eq. (6) indicates that the two particle species would segregate even without the presence of shear, which, of course, does not occur. There are several ways to rationalize this. For example, Gera et al. [44] extended the KTGF drag model to dense uidized beds by adding a \hindrance e ect" term related to the friction and pressure, such that the momentum exchange has to exceed a threshold to drive the segregation. Here we take a heuristic approach by introducing the inertial number I from the (I ) rheology [45{47] to account for local ow conditions, such that bulk I = _d ; (8) where the bulk density is  =   and _ is the local shear rate. A small value of I (small _ and/or large bulk solid solid P ) corresponds to the quasistatic regime where granular ows behave more like deforming solids. Conversely, a large value of I (large _ and/or small P ) corresponds to the collisional regime where granular ows behave more like uids. It is reasonable to expect that the interspecies drag increases with pressure, so the inertial number appears as 1=I in Eq. (7). Finally, the fourth term in Eq. (7) is an empirical modi cation to account for the nonlinear dependence of segregation on particle concentration (i.e., heavy particles among many light particles segregate faster than light particles among many heavy particles [20]), as it increases the drag with increasing  = . h l The ultimate goal of this paper is to propose an accurate model for the segregation velocities of the light and heavy particle species in density bidisperse granular ows. Such an expression can be readily derived by substituting the expression for the interspecies drag from Eq. (7) into the equilibrium momentum balance of Eq. (3). For light particles, interspecies drag , Eq. (7) z }| { h l l p;l h solid 2 (  )g B() ( ) = 0; (9) solid l l | {z } ( +  ) d I l h h l net buoyant forces such that w takes the form p;l 1=2 gd 1 c w = R ) (1 c )I; (10) p;l  l B() R c solid  h noting that  = = c =c and that R =  = is the density ratio. For heavy particles, a similar approach yields h l h l  h l 1=2 gd 1 c w = R ) (1 c )I: (11) p;h  h B() R c solid  h 4 Equations (10) and (11) indicate that w / R , which is similar to the empirical relation w / ( R p;i  p;i ) observed in monodisperse systems with several heavy intruder particles [18]. Equations (10) and (11) also resemble the general form of the segregation velocity given by Fry et al. [17] w = gdf (R )(1 c )I; (12) p;i  i by making 1=2 1 1 c f (R ) = R : (13) B() R c solid  h These similarities suggest that the drag model proposed in Eq. (7) is reasonable and consistent with previous research. III. CONFIRMING THE MODEL The challenge now is to con rm the interspecies drag model in Eq. (7) and validate the resulting segregation model in Eqs. (10) and (11). This is accomplished by performing DEM simulations under a variety of conditions. Direct measurement of from DEM simulations is dicult since both the segregation driving forces and the drag forces result from interparticle collisional forces; it is hard to distinguish them at the particle level. A better way to determine is to utilize momentum conservation at force equilibrium, Eq. (3), in which equals (  ) g, a quantity i i solid i i that is easily measured locally in a granular ow. This is equivalent to validating the segregation model in Eqs. (10) and (11), noting that all of the variables in these equations except B() are known (R ,  , d) or can be calculated solid from the simulation (w , I ). Hence, the problem comprises nding B() under a variety of local ow conditions, I , p;i and local concentration ratios, c =c . h l In the DEM simulations a density bidisperse mixture of d = 4 mm spherical particles with collision time t = 1:25  10 s is sheared between top and bottom horizontal frictional planes in a domain with periodic boundary conditions in the streamwise (x) and spanwise (y) directions as shown in Fig. 1. The massive planar top wall moves horizontally with constant velocity U and is free to move vertically. The position of the top wall is determined by the wall weight and contact forces from the top layer of particles. The distance between the top and bottom walls h remains relatively constant ( uctuating by 2%) after an initial rapid dilation of the particles at ow onset. The domain extends about 70d, or 0:28 m, in the streamwise direction and 10d, or 0:04 m, in the spanwise direction. The weight of the massive planar top wall is based on the con guration of the system so that the overburden pressure applied by the wall to the system is P = 0:05 gh for all cases in this study. There are 36864 particles in wall bulk the system, di erentiated by their densities ( for heavy particles and  for light particles having density ratio h l R =  = ). The initial dense packing is achieved by placing particles in a grid pattern and letting them settle under h l gravity. The depth of the ow h in the vertical z-direction is approximately 50d or 0:2 m, and the volume fraction ratio ( = ) or, equivalently, the concentration ratio (c =c ), is approximately uniform across the domain. Varying h l h l c =c or  = in the simulations causes the segregation velocity to vary. Over 200 simulations are performed over the h l h l wide range of ow conditions and particle properties listed in Table I. To achieve speci c shear rate pro les, the corresponding velocity pro le is imposed on the particle by applying a streamwise stabilizing force on each particle k at every time step according to F = K [u(z ) u ]; (14) stabilize;k s k k where u(z) is the imposed velocity pro le, u is the particle's streamwise velocity, z is the particle's vertical position, k k and K is a gain parameter. The details of the approach for the imposed velocity pro le are provided elsewhere [17]. 2:3(z=h1) 2 2 Three di erent pro les are considered: u = Uz=h, Ue , and Uz =h . The rst and second pro les correspond to those for uniform shear [17] and free surface ow down a heap [48], respectively. The third pro le re ects a situation that sometimes occurs in shear ows where the shear rate vanishes at the bottom wall due to the rough boundary condition [49]. The simulation domain is divided into 20 horizontal layers for averaging purposes; each layer is 2:5d in the z- direction. The streamwise mean velocity pro le for a density segregation simulation 0:1 s after ow onset is shown in Fig. 2 for each of the three imposed velocity pro les. Due to the stabilizing force, the mean velocity matches the imposed pro le, with a shear rate that is either uniform ( _ = U=h) or decreases through the depth of the particle 2 2:3(z=h1) bed [ _ = 2Uz=h , 2:3Ue =h]. The di erent symbols in Fig. 2 represent heavy and light particles, which have similar streamwise velocities. Slight deviations from the imposed velocity pro les occur within 5d from the top and 10d 𝑢 𝑧 Light particles 𝑤 %,' Heavy particles 𝑤 %,( FIG. 1: Schematic of DEM simulation setup. 0.8 0.6 0.4 0.2 0 0.5 1 FIG. 2: Instantaneous streamwise velocity u of each constituent for mixed light (+) and heavy (#) particles averaged in each horizontal layer 0:1 s after shear onset. Dashed curves represent the velocity pro les imposed using the forcing speci ed in Eq. (14). (d = d = 4 mm, c = c = 0:5,  = 0:2, e = 0:9, and R = 4.) h l h l bottom boundaries due to particle ordering adjacent to the at walls [49, 50]. Here, we focus on the ow away from the walls to avoid any artifacts related to these bounding walls. Companion simulations with bumpy walls (particles attached randomly on at walls [51]) show little di erence with simulations using at walls for the particles between 0:3  z=h  0:7. To characterize the evolution of the segregation, we measure the average center of mass height for each species relative to the mean height of all particles, which is calculated as i N X X 1 1 z  = z z ; (15) i k k N N k2i k=1 where N and N are the number of particles of species i and the total number of particles in the horizontal averaging 70d 50d 6 TABLE I: Simulation parameter ranges. d 4 mm R =  = 1 10 h l 0 0:6 c , c 0:1 0:9 h l e 0:2 0:9 U 1 5 m/s 2 2 2:3(z=h1) u Uz=h, Uz =h , Ue 2 2:3(z=h1) _ U=h, 2Uz=h , 2:3Ue =h 0.5 1 0.8 0.6 0.4 0.2 -0.5 0 0 0.5 1 0 2 4 6 FIG. 3: Determining the segregation velocity. (a) Average center of mass displacement z =d time series at di erent vertical positions under uniform shear ( _ = U=h = 25 s and R = 8). (b) Segregation velocity pro les for light particles w () and heavy particles w (#) are calculated as the average rate of change of z  over the rst 1 s of p;l p;h the simulation. Both shear rate _ and density ratio R a ect the segregation velocity pro le. Error bars represent the uncertainty in determining z  and w ; c = c = 0:5,  = 0:2, e = 0:9. p;h h l layer, respectively. Figure 3(a) shows the o set of the center of mass from its initial position for both heavy and light particles in horizontal layers at three vertical locations, z=h = 0:3, 0:5, and 0:7, for a uniform shear ow (u = Uz=h). The o set is represented by the segregation distance z  = z  z  , where z  is the initial center of mass. As the i i i;0 i;0 segregation progresses from shear onset, heavy particles move downward while light particles move upward. Due to solid volume conservation, z  + z  = 0. Since the shear rate is uniform across the domain for the simulation data l h shown in Fig. 3, the di erent segregation rates in the three di erent layers are a consequence of the overburden, or lithostatic, pressure [17]. A higher overburden pressure, which corresponds here to deeper bed depth (small z=h), reduces the segregation. Di erent segregation rates at di erent vertical positions are also found for ows with the nonlinear velocity pro les. However, in such cases, the depth-varying segregation rate results from a combination of shear rate and overburden pressure, since the shear rate now also varies with depth. Con ned granular mixtures under shear (Fig. 1) initially experience a rapidly segregating stage when the segregation velocity remains nearly constant [17]. This is because for a short period after the ow is initiated, the local particle concentration has not yet changed enough to a ect the segregation, resulting in a relatively high segregation velocity, while at later times the particles reach a steady segregated state where the segregation velocity is negligible. We focus on the initial rapidly segregating transient. The shear rate is chosen so that the initial segregation velocity is slow enough that the 1 s sampling window used here falls within the interval of rapid segregation ( 10 s). The linear pro le of the segregation distance in Fig. 3(a) indicates the minimal e ect of the concentration change (less than 5% at most over the sampling window) on the segregation velocity. The segregation velocity of each constituent is calculated from the slope of data like that in Fig. 3(a) as w = (z    )=t, where z  is the o set at the end p;i i;1 i;1 i;1 of the t = 1 s sampling window and  is the standard error in the calculation of z  , which is represented by i;1 i;1 7 the error bars at time t slightly less than 1 s in Fig. 3(a). By considering many thin horizontal layers, the local segregation velocity pro le can be measured, as shown in Fig. 3(b) for three sample cases of uniform shear ow. The segregation velocities increase with both increasing density ratio and shear rate. For each case, segregation velocity decreases from top to bottom even though the shear rate and density ratio do not vary with depth. This is a consequence of the overburden pressure P increasing with depth, consistent with previous results [17]. Note that the sum of segregation velocities of heavy and light particles should equal zero due to volume conservation for c = c = 0:5. Figure 3(b) indeed demonstrates that the segregation h l velocities are nearly equal (and, of course, opposite). The slight di erence between w and w is likely caused p;l p;h by the small amount of segregation during the initial lling process before the start of the simulation. The error bars represent uncertainties in the measurement of the slope in Fig. 3(a). @ 2 From the momentum balance equation, Eq. (2), the convection term in the z-direction (  w ), which can be i i @z @P estimated from the simulation results, is three orders of magnitude less than the net buoyant force c   g i i i @z (noting that the constituent velocity w is equal to the segregation velocity w since the bulk velocity in the z- i p;i direction is zero). Likewise, the unsteady term (  w ) is also negligible. This con rms that the system is in i i i @t steady state during the sampling window and that the segregation velocity results from the balance between the net buoyant force and the interspecies drag as indicated by the simpli ed momentum expression in Eq. (3). A. Inertial number dependence The inertial number dependence assumed in the proposed drag model, Eq. (7), requires that the segregation velocity is linear in I [Eqs. (10) and (11)]. Therefore, to verify the drag model, segregation velocities of light particles w p;l at di erent depths are plotted versus I for di erent density ratios R in Fig. 4 for ows with nonlinear streamwise velocity pro les. Similar results are also found for the segregation velocity of heavy particles w since w  w for p;h p;h p;l ows with c = c = 0:5. The linear relation between w and I in Fig. 4 for varying _ and P (expressed in terms of h l p;i varying I ) con rms the assumed dependence on I and is consistent with results from previous studies [17, 18]. That is, the segregation velocity depends linearly on I for each density ratio. The error bars are somewhat large, consistent with the diculty in accurately measuring w as indicated in Fig. 3. Nevertheless, a line can be t through the data p;i at each density ratio. Furthermore, the tting lines are quite similar for both nonlinear velocity pro les. The data, however, do not show as clear a linear dependence on I when plotted in the same way for uniform shear 2 2 ows, as illustrated in Fig. 5(a). This is because unlike the other two nonlinear velocity pro les [u = Uz =h and 2:3(z=h1) Ue ], the local inertial number in a uniform shear ow is not equal to zero or nearly zero at the bottom wall. Instead, in the case of the linear velocity pro le (u = Uz=h), the segregation is forced to cease at the xed bottom wall because of the wall itself. To account for the cessation of density segregation at the bottom wall, we use a modi ed inertial number for uniform shear ows I = I I ; (16) where I is the inertial number at which density segregation e ectively ceases. According to Fig. 3(b), segregation ceases at z=h  0. Therefore, we assume I is the inertial number at that location, such that I = _d  =P 0 0 bulk 0 with P =  gh + P . As shown in Fig. 5(b), using I instead of I collapses the data for uniform shear ows 0 bulk wall such that the dependence of w on I is close to linear. Note that for the two nonlinear velocity pro les I equals I p;i since I  0. From here on, unless otherwise speci ed, we drop the asterisk on I , using the corrected inertial number [Eq. (16)]. Figures 4 and 5(b) show that w depends linearly on I for varying shear pro les, shear rates, and density ratios. p;i This con rms / I in Eq. (7) since equating the buoyant and drag forces gives w / I in Eqs. (10) and (11). The i p;i linear relation between w and I is also consistent with previous results in free surface ows. Consider, for example, p;i heap ows, in which the owing layer thickness is about eight particle diameters [48, 52]. For such a thin layer with the local shear rate decreasing exponentially with depth, the overburden, or lithostatic, pressure can be treated as a uniform parameter, and the linear relation between w and I can be reduced to a linear relation between w and _d p;i p;i [53, 54]. On the other hand, for a thick owing layer with a uniform shear rate, the segregation velocity at di erent depths in the owing layer is only a ected by the overburden pressure so long as we consider segregation away from the upper and lower bounding walls. The e ect of lithostatic pressure has to be included, resulting in a dependence of w on I [17, 18]. We further note that in our previous work [17], we proposed a form for I in terms of a critical p;i pressure that is equivalent to Eq. (16). However, it should be noted that at larger inertial numbers (I > 0:5) [18], instantaneous binary collisions dominate over enduring contacts, resulting in a rheological change from dense to rapid dilute granular ow. In the latter regime, the linear relation between w and I may no longer be valid. p;i 8 0 0.05 0.1 0.15 0 0.05 0.1 0.15 FIG. 4: Segregation velocities of light particles vs. local inertial number I at di erent depths for nonlinear velocity 2 2 2:3(z=h1) pro les (a) u = Uz =h and (b) u = Ue for simulations with U = 2 m/s, R =  = 2 f2; 4; 10g, h l c = c = 0:5,  = 0:2, and e = 0:9. Dashed lines are linear ts for each density ratio, demonstrating a linear h l dependence of w on I . Error bars represent the uncertainty in determining w as indicated in Fig. 3. p;l p;l 0 0.05 0.1 0.15 0 0.05 0.1 0.15 FIG. 5: Segregation velocities of light particles w at di erent depths vs. (a) inertial number I and (b) modi ed p;l 2 2 inertial number I = I I for simulations with particle properties given in Fig. 4 but for uniform shear rates 1 1 1 _ = 5 s (#), 15 s (2), and 25 s (3). I is the cut-o inertial number at which density segregation e ectively ceases [17]. The dashed line in both gures is identical and a linear t to the data in (b). Error bars represent the uncertainty in determining w as indicated in Fig. 3. p;l B. Density ratio dependence As shown in Fig. 4, w =I increases with density ratio R . Equations (10) and (11) indicate that w =I is proportional p  p to R 1=R . To test this proposed dependence, w =I , calculated as the slope of the tting lines in Figs. 4 and 5(b), p;l is plotted versus R in Fig. 6. The dashed curve in Fig. 6 represents the predictions of Eq. (10) with B( = 0:2) = 400. (We characterize the dependence of B on  in Section III.D.) Indeed, w =I is proportional to R 1=R over the p;l wide range of R tested (1:3  R  10) in di erent ow pro les where I varies due to changes in both P and _ . Good agreement between the curve and the data con rms the proposed dependence of the drag model on particle densities and, equivalently, the functional dependence of w on R in Eqs. (10) and (11). p;i  9 0 5 10 FIG. 6: w =I vs. R for simulations with 1:3  R  10,  = 0:2, e = 0:9, and c = c = 0:5. Colors and symbols p;l   h l 2 2:3(z=h1) represent simulations with _ = U=h (#), 2Uz=h (), and 2:3Ue =h (). The dashed curve shows the dependence of w =I on R from Eq. (10) with B() = 400. p;i C. Concentration ratio dependence A recent study [20] indicates that the segregation in mixtures of heavy and light particles has an underlying asymmetry that depends on the local particle concentration, similar to mixtures where particles di er in size [19]. Speci cally, a heavy particle among mostly light particles segregates faster than a light particle among mostly heavy particles. The last term  = , on the right hand side of the drag model in Eq. (7) accounts for the nonlinear h l dependence of segregation velocity on particle concentration. To con rm this term, we rewrite Eq. (10) using w = p;l c (w w ) from Eq. (5), such that c on the l.h.s and (1 c ) on the r.h.s cancel and the equation becomes h l h h l 1 gd 1 c 2 2 (w w ) = (R )I : (17) l h B()  R c solid  h According to Eq. (17), (w w ) depends linearly on c =c . To test this dependence, we keep the particle densities l h l h and other simulation parameters constant, but vary c from 0.1 to 0.9 in steps of 0.1 for uniform shear ow. Thus, c =c varies from 1=9 to 9. The particle concentration is limited to the range 0.1 to 0.9 because (w w ) is close to l h l h zero for c > 0:9, making the segregation velocity too small to accurately determine. Figure 7 plots the square of the 2 w w 2 solid 1 1 l h normalized velocity di erence between the two species w ^ = ( ) (R ) , excluding B() because I gd R , and hence B(), is the same for all simulations included in the gure, as a function of c =c for uniform shear l h ows. Note that (w w )=I is calculated using the slope of the tting line through the data at di erent depths p;l p;h like that in Fig. 5(b). This avoids propagating the random error evident in these gures as the deviation of the data points for individual simulations from the tting line. By varying only the concentration ratio in Fig. 7, it is evident p p that w ^ depends linearly on c =c . As expected, larger uncertainties exist for small c =c due to the diculty l h l h in measuring small values of the segregation velocity. Nonetheless, Fig. 7 con rms the linear relation between w ^ and c =c used in Eqs. (10) and (11), and thus the proposed form of the drag model [Eq. (7)]. l h D. Friction and restitution dependence The functional form for B() characterizing the dependence of the drag on particle friction, in Eqs. (7), (10), and (11), is dicult to derive analytically due to the complex particle interactions typical of dense granular ows. Instead, DEM simulations under a variety of ow conditions are used to determine B() and demonstrate that it is independent of the restitution coecient e. Consider rst e, which characterizes energy dissipation in short-duration collisions. We vary e from 0.1 to 0.9 for uniform shear ow while keeping all other parameters unchanged, such that the damping coecient of the linear- spring-dashpot collision model used in the DEM simulations varies accordingly. The segregation velocities for both light and heavy particles at z=h = 0:5 for the case with c = c = 0:5 and R = 8 are plotted versus e in Fig. 8(a) for h l  10 0.01 0.005 0 1 2 3 FIG. 7: Square of the normalized velocity di erence w ^ versus square root of concentration ratio c =c for l h simulations with R = 4,  = 0:2, e = 0:9, and _ = U=h = 25 s . c varies from 0.1 to 0.9 in steps of 0.1 and c = 1 c . The dashed line is a linear t to the data forced through zero. l h 0 0.5 1 0 0.3 0.6 FIG. 8: Dependence of segregation velocities w on contact parameters. (a) Light () and heavy (#) particle p;i segregation velocities are nearly independent of restitution coecient e at z=h = 0:5 with  = 0 (black), 0:1 (red), and 0:4 (blue). (b) Particle segregation velocity decreases as  increases for z=h = 0:3 (red), 0:5 (blue), 0:7 (black) with e = 0:9. R = 8, c = c = 0:5, and _ = U=h = 25 s . h l three di erent friction coecients . Although there is some scatter in the data, it is clear that w is independent of p;i e, regardless of , as is reasonable to expect for dense granular mixtures in which short-duration collisions are unlikely to play as important of a role as in dilute ows. For dense granular mixtures, enduring contacts are dominant, and stresses are generated by long-duration sliding and rolling contacts. Unlike e, the surface friction coecient  has signi cant impact on w as shown in Fig. 8(b). The p;i particle segregation velocity generally decreases as  increases at di erent bed depths. Note that w measured from p;i DEM simulations is a ected by not only  but also the  dependent initial packing, which adds more uncertainties resulting in scatter in the data. Nevertheless, the overall trend of decreasing segregation velocity with increasing friction remains evident. In this context, B() is determined through a series of DEM simulations for uniform shear ow with di erent density ratios R and friction coecients . Since the dependence of drag on the other parameters has been veri ed, it is possible to determine the functional form for B() from Eq. (10), 11 0.2 0.15 0.1 0.05 0 100 0 2 4 6 8 0 0.2 0.4 0.6 gd c 1 l 2 2 FIG. 9: (a) (R ) (1 c ) vs. (w =I ) for uniform shear ows with R 2 f2; 4; 6; 8; 10g and three l p;l R c solid  h di erent  values. B() is determined by the tted slope of the dashed lines, which are forced through zero (c = c = 0:5, _ = 25, and e = 0:9). (b) B() vs.  based on 135 simulations. Di erent symbols correspond to h l e = 0:2 (), 0.8 (), and 0.9 (#) for uniform shear ows. gd 1 c w l p;l B() = (R ) (1 c ) : (18) R c I solid  h In applying Eq. (18), the slope of the tting line in Fig. 5(b) is again used for w =I . We estimate B() by plotting p;l gd c 1 l 2 2 (R ) (1 c ) versus (w =I ) over a range of density ratios for three values of  in Fig. 9(a). The l p;l R c solid  h slope of the data at each  gives the corresponding value of B(). By measuring the slopes of di erent sets of data with 0    0:6, we empirically determine the dependence of B() on . The results are plotted in Fig. 9(b) for several values of e for uniform shear ow. For bidisperse mixtures with a small friction coecient ( < 0:3), B() is linearly proportional to . For larger values of  (0:3    0:6), the relation for B() remains linear but with a smaller slope. There is slightly more scatter for di erent values of e at larger . Note that B( = 0:2) = 400 from Fig. 9(b) is consistent with the value used in Fig. 6 to calculate w =I . p;i The interspecies drag is a combination of normal and tangential forces resulting from long-duration contacts. When = 0, the tangential forces vanish. Thus, B( = 0) 6= 0 means that interspecies drag of smooth particle systems is entirely due to normal forces. Figure 9(b) shows that B() increases with . That is, increasing friction increases the interspecies drag, thereby reducing the segregation velocity. This is contrary to observations from size segregation, where part of the segregation driving force comes from interparticle friction, such that increasing friction coecient promotes the segregation of large particles [55]. A possible explanation is that the friction force adds to the segregation driving force in size segregation, which overcomes the resultant increase in the interspecies drag. Alternately, for particles having di erent densities but the same size, the friction force has no impact on the segregation driving force but only adds to the drag. Note that Eqs. (10) and (11) are not limited to speci c velocity pro les even though B() is empirically determined using the data from uniform shear ows with c = c = 0:5. B() is merely a coecient to account for the particle h l frictional properties. Thus, Eqs. (10) and (11) are generally applicable to predict density segregation in any dense segregating granular ow, as shown in the next section. IV. SEGREGATION VELOCITY MODEL Now that the various terms of the segregation velocity model have been validated, it is possible to compare the predictions of Eqs. (10) and (11) to the segregation velocity measured in the DEM simulations. To do this, the global values for d,  , B(), c , c , and R are used in the model along with local values of I at each horizontal position solid l h  12 0.8 0.6 0.4 0.2 0 2 4 0 2 4 FIG. 10: Predictions of segregation velocity pro les from Eqs. (10) and (11) (dashed curves) compared to DEM simulation data for light () and heavy (#) particles under (a) uniform ( _ = U=h) and (b) varying shear rate 2 2:3(z=h1) pro les [ _ =2Uz=h , 2:3Ue =h] with di erent density ratios.  = 0:2, c = c = 0:5, and e = 0:9. h l 0.8 0.6 0.4 0.2 0 2 4 6 8 0 2 4 6 8 FIG. 11: Predictions of segregation velocity pro les from Eqs. (10) and (11) (dashed curves) compared to DEM simulation data under uniform shear for (a) light and (b) heavy particles with heavy particle concentration c =0.3, 0.5 and 0.7 (c = 1 c ).  = 0:2, R = 8, _ = 25 s , and e = 0:9. l h in the ow. The local inertial number I is based on the overburden pressure estimated as P = P +  g(h z), wall bulk such that I can be expressed as a function of vertical position z according to Eq. (8) (or, equivalently, Eq. (16) for uniform shear ows). Figure 10(a) compares the particle segregation velocity pro les predicted by Eqs. (10) and (11), with those measured from DEM simulations of uniform shear ows in Fig. 3(b). Since data for uniform shear ows is used in determing B(), good agreement between the predicted segregation velocities and those from uniform shear ows is expected. However, the segregation velocities for z=h < 0:2 and z=h > 0:8 in the simulations are a ected by the segregating particles accumulating near the bounding walls. 2 2:3(z=h1) A more rigorous test of the model is to consider ows with varying shear rates [ _ = 2Uz=h , 2:3Ue =h] in Fig. 10(b). The close match between the prediction and the data demonstrates that Eqs. (10) and (11) are also e ective for ow with non-uniform shear rates, even though B() in the model was derived from ow with uniform shear rate ( _ = U=h). Figure 11 provides a similar comparison where the concentration ratio c =c is varied while keeping other parameters h l constant for uniform shear. In this case, the segregation velocities for heavy and light particles di er from one another when the concentration of the two species is not equal, as expected [20]. Again, the estimated segregation velocities 13 match the DEM data. Thus, we conclude that the segregation model [Eqs. (10) and (11)] based on the interspecies drag term derived from analogy with KTGF can be used to estimate the segregation velocities of both light and heavy species through the depth of the owing layer under a variety of ow conditions. Unlike the con ned shear ows discussed to this point, the shear rate and the concentration ratio of many free surface ows vary across the domain. To test the model under these more general conditions, we consider quasi-2d bounded one-sided heap ow in which particles ow in a thin surface layer down a slope much like what occurs when lling a silo [53, 56{58]. Figure 12 shows segregation velocities from heap ow DEM simulations [52] versus 1=4 (c =c ) (1 c )I , since, according to Eqs. (10) and (11), these two variables are linearly related when all other l h i conditions are equal. The di erence between these results and those up to this point is that these results correspond to a wide range of shear rates and concentrations, all occurring simultaneously in the owing layer of the heap. The solid lines in Fig. 12 are based on the model, the data points represent DEM results [52], and the dashed lines represent least squares ts through the data points. Overall, the segregation velocities of both bounded heap ows are reasonably well predicted by the model. However, since segregation in heap ows mainly occurs in a thin layer at the free surface, the averages are more uncertain, so the data have substantial scatter, especially for the weakest segregation case with R = 1:84 and d = 3 mm in Fig. 12(a). The case with a small particle diameter d = 2 mm but large density ratio R = 3:3 has overall larger w such that the DEM results match the theoretical predictions p;i better, as shown in Fig. 12(b). Note that we only include the heap ow data for 0:1  c ; c  0:9 to match the h l concentration range for ows upon which our model is based. As Fig. 12 shows, the segregation model we propose in this paper works reasonably well not only for the con ned shear ow between two planes from which it is developed, but also for free surface ow in a bounded heap. Previous results [52] suggest that the segregation velocity can be modeled as jw j = S _ (1 c ) (19) p;i D i for bounded heap ow with inlet concentration c = c = 0:5. Equations (10) and (11) have this form when S , the l h D segregation length scale, is 1=2 S 1 1 c  gd D l bulk = R : (20) d B() R c P solid  h Previous results for free surface ows [52] indicate that the empirically determined segregation length scale S is well approximated by = C lnR ; (21) where C = 0:081. The empirical relation and the underlying data are shown in Fig. 13(a). To compare S and S it is necessary to account for the pressure and particle concentration ratio, as well as D D other ow conditions. The comparison is made by rst assuming c =c = 1, which is the feed concentration ratio in l h these simulations, such that S is no longer concentration dependent. Furthermore, in heap ows, the segregation occurs in a thin owing layer with a thickness of only a few particle diameters, in which the local shear rate decreases exponentially with depth while the scaled local pressure P=( gd) increases linearly from 1 at the surface to =d at bulk the bottom of the owing layer at depth , assuming constant  . The mean shear rate of the owing layer equals bulk the local shear rate at a depth of 0:26 due to the exponential streamwise velocity pro le. This depth corresponds to approximately 2d for 7d    8:5d [52]. By assuming the exponentially varying local shear rate is the dominant factor for the depth varying segregation velocity, we expect that the e ective pressure is equivalent to the pressure at a depth of about 2d such that P=( gd) = 2. For comparison, P=( gd) = 1 and 3 are also considered. bulk bulk The expression for S from Eq. (20) for c =c = 1 and three di erent pressures is compared to bounded heap ow D h l results [52] in Fig. 13(b). The data match the model curve well for P=( gh) = 2. Again, the results in Fig. 13(b) bulk demonstrate the success of the model in replicating a wide range of results even down to the slight curvature evident in the data matching that of the curve corresponding to Eq. (21). 14 light light -2 heavy heavy -4 0 0.04 0.08 0.12 0.16 0 0.04 0.08 0.12 0.16 FIG. 12: Segregation velocities from DEM heap ow simulations (data points) [52]. (a) R = 1:84,  = 0:2, e = 0:9, and d = 3 mm. (b) R = 3:33,  = 0:2, e = 0:9, and d = 2 mm. Dashed lines are linear ts to the data points, and solid lines are predictions of the model. 0.3 0.3 0.2 0.2 0.1 0.1 0 0 -0.1 -0.1 -0.2 -0.2 -0.3 -0.3 -1 1 -1 1 10 1 10 10 10 FIG. 13: Dependence of segregation length scale S on the density ratio R from DEM simulations (circles) for segregation in a bounded heap ow [52] compared to predictions of (a) the empirical model Eq. (21) and (b) Eq. (20) derived from the segregation velocity model with c =c = 1,  = 0:2 and di erent values of P=( gd). h l bulk V. CONCLUSION In this paper, we have proposed a predictive model, Eqs. (10) and (11), for the segregation velocities of light and heavy particle species in density bidisperse granular ows that is based on a new model for the interspecies drag in segregating dense ows, Eq. (7). The interspecies drag model assumes that the multiple long-duration particle interactions in dense granular ows re ect similar physics to short-duration binary particle interactions typical of dilute granular ows, which have been successfully modeled using the Kinetic Theory of Granular Flows (KTGF) [29, 30]. Of course, particle segregation depends on the pressure-shear state, which can be characterized by the inertial number I . In particular, I is inversely proportional to the square root of the overburden pressure, which can signi cantly reduce the segregation velocity [17]. In addition, the segregation depends non-linearly on the local particle concentration: heavy particles among many light particles segregate more quickly than light particles among many heavy particles [20]. Finally, particles in dense granular ows experience enduring contacts characterized by interparticle friction, which is in contrast to dilute granular ows where particle contacts are short and dominated by the elastic properties of the particles. Hence, the restitution coecient e has little in uence on segregation in the 15 dense ows considered here. The advantage of the proposed interspecies drag model for dense granular ows [Eq. (7)] is that it links interspecies drag to segregation velocities and includes the e ects of the local ow condition, I , local concentration ratio, c =c , h l and particle properties including the particle densities, sizes, and surface friction, . The segregation velocities derived from combining the interspecies drag model with the equilibrium momentum balance equation match the segregation velocities determined from DEM simulations in both con ned shear and free surface heap ows. This allows calculation of segregation velocities through the depth of the owing layer for di erent density ratios and relative constituent concentrations. Despite these advances, the modi ed KTGF drag model proposed here is not without drawbacks. The segregation velocities predicted from Eqs. (10) and (11), while showing the right trends, are only reasonable estimates, as is evident from the scatter in the DEM simulation data from the model predictions in Figs. 10-12. This is likely a result of the many variables in the problem, as well as the stochastic nature of the forces on individual particles that drive the segregation and the concurrent collisional di usion. Nevertheless, the results in Fig. 13 demonstrate a remarkable correspondence between the segregation velocity model and the simulation data, particularly since the data are for a di erent ow than that from which the model was derived. More research is needed under even more widely varying conditions to re ne the model. Furthermore, the model is limited to density bidisperse granular materials. A more challenging problem is to connect the interspecies drag to the segregation velocity for size-bidisperse particles, or, even more dicult, particles that di er in both size and density. [1] J. M. N. T. Gray and A. R. Thornton, A theory for particle size segregation in shallow granular free-surface ows, Proc. R. Soc. A 461, 1447 (2005). [2] S. W. Meier, R. M. Lueptow, and J. M. Ottino, A dynamical systems approach to mixing and segregation of granular materials in tumblers, Adv. Phys. 56, 757 (2007). [3] Y. Fan and K. M. Hill, Theory for shear-induced segregation of dense granular mixtures, New J. Phys. 13 (2011). [4] D. R. Tunuguntla, O. Bokhove, and A. R. Thornton, A mixture theory for size and density segregation in shallow granular free-surface ows, J. Fluid Mech. 749, 99 (2014). [5] H. Xiao, D. McDonald, Y. Fan, P. B. Umbanhowar, J. M. Ottino, and R. M. Lueptow, Controlling granular segregation using modulated ow, Powder Technol. 312, 360 (2017). [6] Y. Zhao, H. Xiao, P. B. Umbanhowar, and R. M. Lueptow, Simulation and modeling of segregating rods in quasi-2d bounded heap ow, AIChE J. 64, 1550 (2018). [7] H. Xiao, Y. Fan, K. V. Jacob, P. B. Umbanhowar, M. Kodam, J. F. Koch, and R. M. Lueptow, Continuum modeling of granular segregation during hopper discharge, Chem. Eng. Sci. 193, 188 (2019). [8] P. B. Umbanhowar, R. M. Lueptow, and J. M. Ottino, Modeling segregation in granular ows, Annu. Rev. Chem. Biomol. Eng. 10 (2019). [9] D. V. Khakhar, J. J. McCarthy, and J. M. Ottino, Radial segregation of granular mixtures in rotating cylinders, Phys. Fluids 9, 3600 (1997). [10] D. V. Khakhar, J. J. McCarthy, and J. M. Ottino, Mixing and segregation of granular materials in chute ows, Chaos 9, 594 (1999). [11] A. Tripathi and D. V. Khakhar, Density di erence-driven segregation in a dense granular ow, J. Fluid Mech. 717, 643 (2013). [12] P. Gajjar and J. M. N. T. Gray, Asymmetric ux models for particle-size segregation in granular avalanches, J. Fluid Mech. 757, 297 (2014). [13] J. M. N. T. Gray and C. Ancey, Particle-size and-density segregation in granular free-surface ows, J. Fluid Mech. 779, 622 (2015). [14] B. Marks, P. Rognon, and I. Einav, Grainsize dynamics of polydisperse granular segregation down inclined planes, J. Fluid Mech. 690, 499 (2012). [15] J. M. N. T. Gray, Particle segregation in dense granular ows, Annu. Rev. Fluid Mech. 50, 407 (2018). [16] T. Weinhart, S. Luding, and A. R. Thornton, From discrete particles to continuum elds in mixtures, in AIP conference proceedings, Vol. 1542 (AIP, 2013) pp. 1202{1205. [17] A. M. Fry, P. B. Umbanhowar, J. M. Ottino, and R. M. Lueptow, E ect of pressure on segregation in granular shear ows, Phys. Rev. E 97, 062906 (2018). [18] S. Liu and J. J. McCarthy, Transport analogy for segregation and granular rheology, Phys. Rev. E 96, 020901 (2017). [19] K. van der Vaart, P. Gajjar, G. Epely-Chauvin, N. Andreini, J. M. N. T. Gray, and C. Ancey, Underlying asymmetry within particle size segregation, Phys. Rev. Lett. 114, 238001 (2015). [20] R. P. Jones, A. B. Isner, H. Xiao, J. M. Ottino, P. B. Umbanhowar, and R. M. Lueptow, Asymmetric concentration dependence of segregation uxes in granular ows, Phys. Rev. Fluids 3, 094304 (2018). [21] D. Gidaspow, M. Syamlal, and Y. Seo, Hydrodynamics of uidization of single and binary size particles: supercomputer modeling, in Proceedings of Fluidization V (AIChE Engineering Foundation, 1986) pp. 1{8. [22] M. Syamlal, The particle-particle drag term in a multiparticle model of uidization, Tech. Rep. (DOE/MC/21353-2373:Work 16 perfomed under contract No: DE-AC21-85MC21353, 1987). [23] L. Huilin and D. Gidaspow, Hydrodynamics of binary uidization in a riser: CFD simulation using two granular temper- atures, Chem. Eng. Sci. 58, 3777 (2003). [24] L. Huilin, Z. Yunhua, J. Ding, D. Gidaspow, and L. Wei, Investigation of mixing/segregation of mixture particles in gas{solid uidized beds, Chem. Eng. Sci. 62, 301 (2007). [25] H. Iddir, H. Arastoopour, and C. M. Hrenya, Analysis of binary and ternary granular mixtures behavior using the kinetic theory approach, Powder Technol. 151, 117 (2005). [26] H. Iddir and H. Arastoopour, Modeling of multitype particle ow using the kinetic theory approach, AIChE J. 51, 1620 (2005). [27] Z. Chao, Y. Wang, J. P. Jakobsen, M. Fernandino, and H. A. Jakobsen, Investigation of the particle{particle drag in a dense binary uidized bed, Powder Technol. 224, 311 (2012). [28] J. M. N. T. Gray and V. Chugunov, Particle-size segregation and di usive remixing in shallow granular avalanches, J. Fluid Mech. 569, 365 (2006). [29] J. T. Jenkins and S. B. Savage, A theory for the rapid ow of identical, smooth, nearly elastic, spherical particles, J. Fluid Mech. 130, 187 (1983). [30] C. K. K. Lun, S. B. Savage, D. J. Je rey, and N. Chepurniy, Kinetic theories for granular ow: inelastic particles in couette ow and slightly inelastic particles in a general ow eld, J. Fluid Mech. 140, 223 (1984). [31] C. h. Liu, S. R. Nagel, D. A. Schecter, S. N. Coppersmith, S. Majumdar, O. Narayan, and T. A. Witten, Force uctuations in bead packs, Science 269, 513 (1995). [32] J. T. Jenkins, Dense shearing ows of inelastic disks, Phys. Fluids 18, 103307 (2006). [33] J. T. Jenkins, Dense inclined ows of inelastic spheres, Granul. Matter 10, 47 (2007). [34] J. T. Jenkins and D. Berzi, Dense inclined ows of inelastic spheres: tests of an extension of kinetic theory, Granul. Matter 12, 151 (2010). [35] D. Vescovi, D. Berzi, P. Richard, and N. Brodu, Plane shear ows of frictionless spheres: Kinetic theory and 3D soft-sphere discrete element method simulations, Phys. Fluids 26, 053305 (2014). [36] Y. Duan, Z.-G. Feng, E. E. Michaelides, and S. Mao, Modi ed kinetic theory applied to the shear ows of granular materials, Phys. Fluids 29, 043302 (2017). [37] Y. Duan and Z.-G. Feng, A new kinetic theory model of granular ows that incorporates particle sti ness, Phys. Fluids 31, 013301 (2019). [38] M. Larcher and J. T. Jenkins, Segregation and mixture pro les in dense, inclined ows of two types of spheres, Phys. Fluids 25, 113301 (2013). [39] M. Larcher and J. T. Jenkins, The evolution of segregation in dense inclined ows of binary mixtures of spheres, J. Fluid Mech. 782, 405 (2015). [40] M. Syamlal, W. Rogers, and T. J. OBrien, MFIX documentation: Theory guide, National Energy Technology Laboratory, Department of Energy, Technical Note DOE/METC-95/1013 and NTIS/DE95000031 (1993). [41] O. Owoyemi, L. Mazzei, and P. Lettieri, CFD modeling of binary- uidized suspensions and investigation of role of particle{ particle drag on mixing and segregation, AIChE J. 53, 1924 (2007). [42] F. Hern andez-Jim enez, L. Garc a-Guti errez, A. Soria-Verdugo, and A. Acosta-Iborra, Fully coupled TFM-DEM simulations to study the motion of fuel particles in a uidized bed, Chem. Eng. Sci. 134, 57 (2015). [43] J. L. Lebowitz, Exact solution of generalized percus-yevick equation for a mixture of hard spheres, Phys. Rev. 133, A895 (1964). [44] D. Gera, M. Syamlal, and T. J. O'Brien, Hydrodynamics of particle segregation in uidized beds, InterNatl. J. (Wash.) of Multiphase Flow 30, 419 (2004). [45] GDR MiDi, On dense granular ows, Eur. Phys. J. E. 14, 341 (2004). [46] F. da Cruz, S. Emam, M. Prochnow, J.-N. Roux, and F. Chevoir, Rheophysics of dense granular materials: Discrete simulation of plane shear ows, Phys. Rev. E 72, 021309 (2005). [47] P. Jop, Y. Forterre, and O. Pouliquen, A constitutive law for dense granular ows, Nature 441, 727 (2006). [48] Y. Fan, P. B. Umbanhowar, J. M. Ottino, and R. M. Lueptow, Kinematics of monodisperse and bidisperse granular ows in quasi-two-dimensional bounded heaps, Proc. R. Soc. A 469, 20130235 (2013). [49] L. Jing, C. Y. Kwok, Y. F. Leung, and Y. D. Sobral, E ect of geometric base roughness on size segregation, in EPJ Web of Conferences, Vol. 140 (EDP Sciences, 2017) p. 03056. [50] K. van der Vaart, A. R. Thornton, C. G. Johnson, T. Weinhart, L. Jing, P. Gajjar, J. M. N. T. Gray, and C. Ancey, Breaking size-segregation waves and mobility feedback in dense granular avalanches, Granul. Matter 20, 46 (2018). [51] L. Jing, C. Y. Kwok, Y. F. Leung, and Y. D. Sobral, Characterization of base roughness for granular chute ows, Phys. Rev. E 94, 052901 (2016). [52] H. Xiao, P. B. Umbanhowar, J. M. Ottino, and R. M. Lueptow, Modelling density segregation in owing bidisperse granular materials, Proc. R. Soc. A 472, 20150856 (2016). [53] Y. Fan, C. P. Schlick, P. B. Umbanhowar, J. M. Ottino, and R. M. Lueptow, Modelling size segregation of granular materials: the roles of segregation, advection and di usion, J. Fluid Mech. 741, 252 (2014). [54] C. P. Schlick, Y. Fan, A. B. Isner, P. B. Umbanhowar, J. M. Ottino, and R. M. Lueptow, Modeling segregation of bidisperse granular materials using physical control parameters in the quasi-2d bounded heap, AIChE J. 61, 1524 (2015). [55] L. Jing, C. Y. Kwok, and Y. F. Leung, Micromechanical origin of particle size segregation, Phys. Rev. Lett. 118, 118001 (2017). [56] Y. Fan, Y. Boukerkour, T. Blanc, P. B. Umbanhowar, J. M. Ottino, and R. M. Lueptow, Strati cation, segregation, and 17 mixing of granular materials in quasi-two-dimensional bounded heaps, Phys. Rev. E 86, 051305 (2012). [57] Z. Deng, P. B. Umbanhowar, J. M. Ottino, and R. M. Lueptow, Continuum modelling of segregating tridisperse granular chute ow, Proc. R. Soc. A 474, 20170384 (2018). [58] Z. Deng, P. B. Umbanhowar, J. M. Ottino, and R. M. Lueptow, Modeling segregation of polydisperse granular materials in developing and transient free-surface ows, AIChE J. 65, 882 (2019). http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Condensed Matter arXiv (Cornell University)

Kinetic theory based segregation model for density-bidisperse dense granular flows

Free
Download PDF
17 pages

Loading next page...
 
/lp/arxiv-cornell-university/kinetic-theory-based-segregation-model-for-density-bidisperse-dense-jG0q3PetWU
ISSN
2469-990X
eISSN
ARCH-3331
DOI
10.1103/PhysRevFluids.5.044301
Publisher site
See Article on Publisher Site

Abstract

Kinetic theory based segregation model for density-bidisperse dense granular ows 1 2 1, 2, 3 1, 2, 3, Yifei Duan, Paul B. Umbanhowar, Julio M. Ottino, and Richard M. Lueptow Department of Chemical and Biological Engineering, Northwestern University, Evanston, Illinois 60208, USA, Department of Mechanical Engineering, Northwestern University, Evanston, Illinois 60208, USA Northwestern Institute on Complex Systems (NICO), Northwestern University, Evanston, Illinois 60208, USA, (Dated: June 11, 2019) Individual constituent balance equations are often used to derive expressions for species speci c segregation velocities in ows of dense granular mixtures. We propose an expression for the in- terspecies momentum exchange in density bidisperse granular ows as an extension of ideas from kinetic theory, expanding its range from short-duration binary collisions to the multiple enduring contacts characteristic of dense shear ows, and incorporating the e ects of particle friction, con- centration ratio, and local ow conditions. The segregation velocity derived from the momentum balance equation with the new interspecies drag model matches results from DEM simulations and accurately predicts the downward and upward segregation velocities of heavy and light particles through the depth of the owing layer for di erent density ratios and constituent concentrations for both con ned shear and free surface heap ows. I. INTRODUCTION When sheared, dense granular materials tend to segregate by particle size, density, shape or friction coecient, which can be problematic in many industries due to its impact on product quality and uniformity [1{8]. As a result, the segregation behavior of sheared granular mixtures is of interest in a wide variety of elds from both fundamental and applied standpoints. We focus on bidisperse granular mixtures of particles having the same size but di erent densities such that segre- gation is driven by a \buoyant force" mechanism [9, 10]. ln this paper, we propose a model for the interspecies drag between heavy and light particles in density bidisperse granular ows and use the model to derive an expression for the segregation velocity of each species by considering the balance between the net buoyant force and the interspecies drag force. To model the segregation process, we assume that two di erent particle species can be treated as interpenetrable continua such that each constituent has its own partial pressure, consistent with previous studies [1, 3, 4]. The imbalance between a species' partial pressure and its body force results in segregation of the two species. A linear relation between the drag force on a particle and its segregation velocity has been used in several gravity-driven segregation models for size or density segregation [1, 4, 11{14] based on either the similarity of kinetic sieving in granular segregation with uid percolation through a porous material or an analogy to the drag force on a sphere settling in a uid. The linear drag model has been combined with the constituent mass and momentum balances to derive a general multicomponent theory for segregation [15]. However, Weinhart et al. [16] demonstrated that the coecient tted from the linear drag model varies with time, indicating that a simple linear drag law fails to describe the segregation behavior. Recent DEM simulations also suggest that other factors, such as the local pressure [17, 18] and the species concentration ratio [19, 20], a ect the segregation velocity, which again raises questions concerning the validity of a linear drag model. As an alternative to the linear drag model, it is natural to consider the Kinetic Theory of Granular Flow (KTGF) to model the momentum transfer between two segregating species. The KTGF provides a drag model connecting stresses and velocities in a granular mixture while accounting for complex particle interactions [21{27], thereby yielding a drag model that takes particle density and diameter into consideration at the particle level. However, there are two major problems with using the KTGF approach to model dense ows. First, the stress generation mechanism in the KTGF is based on short-duration, binary collisions typical of dilute particle ows as opposed to multiple, enduring contacts typical of the dense granular ows considered here. Second, the KTGF approach does not include physics known to be critical to segregation in dense granular ows such as the dependence of segregation on the local pressure [17]. Nevertheless, a potential path toward a robust drag model for dense ows lies in modifying the KTGF drag model based on empirical results for density segregation in dense granular ows. In this paper, we propose an expression for the density segregation velocity based on mixture theory [1] with a drag model based on KTGF but extended to the dense regime by introducing appropriate coecients. The proposed drag Corresponding author: r-lueptow@northwestern.edu arXiv:1906.03378v1 [cond-mat.soft] 8 Jun 2019 2 model is described in Section II, and an expression for the segregation velocity in density bidisperse granular ow is derived by combining the drag and the equilibrium momentum balance equation. The e ects of the local inertial number, particle friction, and relative species concentration proposed in the model are tested separately by DEM simulations in Section III. In Section IV, results of the segregation model are compared with DEM simulations for ows under more general conditions. Conclusions are given in Section V. II. KTGF BASED DRAG MODEL AND SEGREGATION VELOCITY For dense granular ows, the total solids volume fraction,  =  , where  represents the local volume solid i i fraction of each solid species (here i = h for heavy particles and i = l for light particles), is around 0.6. The local constituent concentration is c =  = and the solid density is  =  c , where  is the particle material i i solid solid i i i density. The gradient of the lithostatic pressure P is @P =   g; (1) solid solid @z where z is the vertical coordinate (assuming negative z is the gravity direction) and g is the acceleration due to gravity. The partial pressure P is determined by partitioning the lithostatic pressure induced by gravity among the two constituents such that P = P . For particles that only di er in density, we assume that the proportion of the hydrostatic load carried by each species equals its concentration c regardless of the density di erence as proposed by Marks et al. [14] and Tunuguntla et al. [4]. The momentum balance for each species can be written as [4] (  uuu ) +rrr (  uuu uuu ) = c rrrP +   ggg + ; (2) i i i i i i i i i i i @t where is the dyadic product, t is time, uuu is the vector velocity, and is the interspecies drag vector. (Eq. (2) is i i similar to Eq. (2.5) in [4], noting that there the intrinsic density  is the bulk density of particles in a non-mixed i i i state, which is equal to   , and that  = c  is equivalent to   used here.) Assuming that the vertical i solid i i acceleration terms are negligible [4, 28], which is reasonable for the relatively slow segregation in a typical granular owing layer, the momentum conservation equation in the z-direction can be simpli ed using Eq. (1) to (  ) g + = 0: (3) solid i i i Equation (3) is analogous to a simple force balance between the the net buoyant force (  ) g and the solid i i interspecies drag force . To determine constituent velocities from Eq. (3), must be expressed as a function of the velocity di erence between the two species, and the KTGF o ers a means to do this. Changes to the KTGF are needed, however, because it was developed to model rapid ow of dilute granular materials based on instantaneous binary collisions [29, 30]. This is not the case in the dense ow regime where enduring contacts dominate and many adjacent particles are part of contact force chains [31]. Correlations in motion and force due to the dense contact network reduce the collisional energy dissipation [32]. Substantial e ort has been expended to extend the applicability of the KTGF to the dense regime, mainly by modifying the expression for the collisional energy dissipation rate to account for the e ects of sustained contacts [33{37]. In fact, an augmented KTGF approach has been proposed for segregation problems, but it is limited to small size and density di erences [38, 39]. In addition, several stand-alone drag models based on the KTGF have been proposed for dilute granular ows [21{27]. Here we follow an approach similar to one of these models [27] by assuming that the long-duration collisional drag follows similar physics to the short-duration collisional drag, but with a much longer contact time. This can be represented by correction coecients modifying the well-known solid-solid drag model proposed by Syamlal [22] that has been successfully used for uidized bed simulations [40{42]. The Syamlal expression for the local drag between two constituents i and j is 3(1 + e) = + g     (d + d ) juuu uuu j(uuu uuu ); (4) i ij i j i j i j i j i j 3 3 2( d +  d ) 2 8 i j i j where g is the radial distribution function for a collision between particle i and j [43],  is the friction coecient, ij and e is the restitution coecient. For density segregation, d = d , and g becomes a constant if we assume i j ij constant total solid volume fraction  . The bulk velocity in the vertical direction of the combined constituents solid 3 is w = c w + c w based on volume conservation. The segregation, or percolation velocity w , for species i is bulk i i j j p;i de ned as the velocity di erence between the constituent velocity and the bulk velocity in the z direction: w = w w = c (w w ): (5) p;i i bulk j i j Using Eq. (5) in Eq. (4) for bidisperse granular mixtures of heavy and light particles having the same diameter d, the interspecies drag force in the segregation (z) direction is jw j h l i p;i = 3(1 + e) 1 + g  w : (6) i ij solid p;i 4 ( +  ) d h l j Based on Eq. (6), we propose a semi-empirical interspecies drag model of the form local concentration ratio KTGF z }| { z }| { jw j 1 h l i p;i h = B()  w : (7) i solid p;i | {z } ( +  ) d I l h j l |{z} enduring contacts local ow Equation (7) has four multiplicative terms. Consider them from left to right. B() replaces the rst term in square brackets in Eq. (6) to account for enduring contacts in dense ows. In dense granular mixtures short-duration collisions are unlikely to play as important of a role as in the dilute ows upon which Eq. (6) for KTGF is based. As a result, we assume B() is independent of e, which is proven in Section III.D. The second term is inherited from the KTGF and re ects the momentum transfer between two particles with di erent mass, taking into consideration density and diameter at the particle level. The third term accounts for the local ow conditions, since Eq. (6) does not explicitly account for the presence of shear. As such, Eq. (6) indicates that the two particle species would segregate even without the presence of shear, which, of course, does not occur. There are several ways to rationalize this. For example, Gera et al. [44] extended the KTGF drag model to dense uidized beds by adding a \hindrance e ect" term related to the friction and pressure, such that the momentum exchange has to exceed a threshold to drive the segregation. Here we take a heuristic approach by introducing the inertial number I from the (I ) rheology [45{47] to account for local ow conditions, such that bulk I = _d ; (8) where the bulk density is  =   and _ is the local shear rate. A small value of I (small _ and/or large bulk solid solid P ) corresponds to the quasistatic regime where granular ows behave more like deforming solids. Conversely, a large value of I (large _ and/or small P ) corresponds to the collisional regime where granular ows behave more like uids. It is reasonable to expect that the interspecies drag increases with pressure, so the inertial number appears as 1=I in Eq. (7). Finally, the fourth term in Eq. (7) is an empirical modi cation to account for the nonlinear dependence of segregation on particle concentration (i.e., heavy particles among many light particles segregate faster than light particles among many heavy particles [20]), as it increases the drag with increasing  = . h l The ultimate goal of this paper is to propose an accurate model for the segregation velocities of the light and heavy particle species in density bidisperse granular ows. Such an expression can be readily derived by substituting the expression for the interspecies drag from Eq. (7) into the equilibrium momentum balance of Eq. (3). For light particles, interspecies drag , Eq. (7) z }| { h l l p;l h solid 2 (  )g B() ( ) = 0; (9) solid l l | {z } ( +  ) d I l h h l net buoyant forces such that w takes the form p;l 1=2 gd 1 c w = R ) (1 c )I; (10) p;l  l B() R c solid  h noting that  = = c =c and that R =  = is the density ratio. For heavy particles, a similar approach yields h l h l  h l 1=2 gd 1 c w = R ) (1 c )I: (11) p;h  h B() R c solid  h 4 Equations (10) and (11) indicate that w / R , which is similar to the empirical relation w / ( R p;i  p;i ) observed in monodisperse systems with several heavy intruder particles [18]. Equations (10) and (11) also resemble the general form of the segregation velocity given by Fry et al. [17] w = gdf (R )(1 c )I; (12) p;i  i by making 1=2 1 1 c f (R ) = R : (13) B() R c solid  h These similarities suggest that the drag model proposed in Eq. (7) is reasonable and consistent with previous research. III. CONFIRMING THE MODEL The challenge now is to con rm the interspecies drag model in Eq. (7) and validate the resulting segregation model in Eqs. (10) and (11). This is accomplished by performing DEM simulations under a variety of conditions. Direct measurement of from DEM simulations is dicult since both the segregation driving forces and the drag forces result from interparticle collisional forces; it is hard to distinguish them at the particle level. A better way to determine is to utilize momentum conservation at force equilibrium, Eq. (3), in which equals (  ) g, a quantity i i solid i i that is easily measured locally in a granular ow. This is equivalent to validating the segregation model in Eqs. (10) and (11), noting that all of the variables in these equations except B() are known (R ,  , d) or can be calculated solid from the simulation (w , I ). Hence, the problem comprises nding B() under a variety of local ow conditions, I , p;i and local concentration ratios, c =c . h l In the DEM simulations a density bidisperse mixture of d = 4 mm spherical particles with collision time t = 1:25  10 s is sheared between top and bottom horizontal frictional planes in a domain with periodic boundary conditions in the streamwise (x) and spanwise (y) directions as shown in Fig. 1. The massive planar top wall moves horizontally with constant velocity U and is free to move vertically. The position of the top wall is determined by the wall weight and contact forces from the top layer of particles. The distance between the top and bottom walls h remains relatively constant ( uctuating by 2%) after an initial rapid dilation of the particles at ow onset. The domain extends about 70d, or 0:28 m, in the streamwise direction and 10d, or 0:04 m, in the spanwise direction. The weight of the massive planar top wall is based on the con guration of the system so that the overburden pressure applied by the wall to the system is P = 0:05 gh for all cases in this study. There are 36864 particles in wall bulk the system, di erentiated by their densities ( for heavy particles and  for light particles having density ratio h l R =  = ). The initial dense packing is achieved by placing particles in a grid pattern and letting them settle under h l gravity. The depth of the ow h in the vertical z-direction is approximately 50d or 0:2 m, and the volume fraction ratio ( = ) or, equivalently, the concentration ratio (c =c ), is approximately uniform across the domain. Varying h l h l c =c or  = in the simulations causes the segregation velocity to vary. Over 200 simulations are performed over the h l h l wide range of ow conditions and particle properties listed in Table I. To achieve speci c shear rate pro les, the corresponding velocity pro le is imposed on the particle by applying a streamwise stabilizing force on each particle k at every time step according to F = K [u(z ) u ]; (14) stabilize;k s k k where u(z) is the imposed velocity pro le, u is the particle's streamwise velocity, z is the particle's vertical position, k k and K is a gain parameter. The details of the approach for the imposed velocity pro le are provided elsewhere [17]. 2:3(z=h1) 2 2 Three di erent pro les are considered: u = Uz=h, Ue , and Uz =h . The rst and second pro les correspond to those for uniform shear [17] and free surface ow down a heap [48], respectively. The third pro le re ects a situation that sometimes occurs in shear ows where the shear rate vanishes at the bottom wall due to the rough boundary condition [49]. The simulation domain is divided into 20 horizontal layers for averaging purposes; each layer is 2:5d in the z- direction. The streamwise mean velocity pro le for a density segregation simulation 0:1 s after ow onset is shown in Fig. 2 for each of the three imposed velocity pro les. Due to the stabilizing force, the mean velocity matches the imposed pro le, with a shear rate that is either uniform ( _ = U=h) or decreases through the depth of the particle 2 2:3(z=h1) bed [ _ = 2Uz=h , 2:3Ue =h]. The di erent symbols in Fig. 2 represent heavy and light particles, which have similar streamwise velocities. Slight deviations from the imposed velocity pro les occur within 5d from the top and 10d 𝑢 𝑧 Light particles 𝑤 %,' Heavy particles 𝑤 %,( FIG. 1: Schematic of DEM simulation setup. 0.8 0.6 0.4 0.2 0 0.5 1 FIG. 2: Instantaneous streamwise velocity u of each constituent for mixed light (+) and heavy (#) particles averaged in each horizontal layer 0:1 s after shear onset. Dashed curves represent the velocity pro les imposed using the forcing speci ed in Eq. (14). (d = d = 4 mm, c = c = 0:5,  = 0:2, e = 0:9, and R = 4.) h l h l bottom boundaries due to particle ordering adjacent to the at walls [49, 50]. Here, we focus on the ow away from the walls to avoid any artifacts related to these bounding walls. Companion simulations with bumpy walls (particles attached randomly on at walls [51]) show little di erence with simulations using at walls for the particles between 0:3  z=h  0:7. To characterize the evolution of the segregation, we measure the average center of mass height for each species relative to the mean height of all particles, which is calculated as i N X X 1 1 z  = z z ; (15) i k k N N k2i k=1 where N and N are the number of particles of species i and the total number of particles in the horizontal averaging 70d 50d 6 TABLE I: Simulation parameter ranges. d 4 mm R =  = 1 10 h l 0 0:6 c , c 0:1 0:9 h l e 0:2 0:9 U 1 5 m/s 2 2 2:3(z=h1) u Uz=h, Uz =h , Ue 2 2:3(z=h1) _ U=h, 2Uz=h , 2:3Ue =h 0.5 1 0.8 0.6 0.4 0.2 -0.5 0 0 0.5 1 0 2 4 6 FIG. 3: Determining the segregation velocity. (a) Average center of mass displacement z =d time series at di erent vertical positions under uniform shear ( _ = U=h = 25 s and R = 8). (b) Segregation velocity pro les for light particles w () and heavy particles w (#) are calculated as the average rate of change of z  over the rst 1 s of p;l p;h the simulation. Both shear rate _ and density ratio R a ect the segregation velocity pro le. Error bars represent the uncertainty in determining z  and w ; c = c = 0:5,  = 0:2, e = 0:9. p;h h l layer, respectively. Figure 3(a) shows the o set of the center of mass from its initial position for both heavy and light particles in horizontal layers at three vertical locations, z=h = 0:3, 0:5, and 0:7, for a uniform shear ow (u = Uz=h). The o set is represented by the segregation distance z  = z  z  , where z  is the initial center of mass. As the i i i;0 i;0 segregation progresses from shear onset, heavy particles move downward while light particles move upward. Due to solid volume conservation, z  + z  = 0. Since the shear rate is uniform across the domain for the simulation data l h shown in Fig. 3, the di erent segregation rates in the three di erent layers are a consequence of the overburden, or lithostatic, pressure [17]. A higher overburden pressure, which corresponds here to deeper bed depth (small z=h), reduces the segregation. Di erent segregation rates at di erent vertical positions are also found for ows with the nonlinear velocity pro les. However, in such cases, the depth-varying segregation rate results from a combination of shear rate and overburden pressure, since the shear rate now also varies with depth. Con ned granular mixtures under shear (Fig. 1) initially experience a rapidly segregating stage when the segregation velocity remains nearly constant [17]. This is because for a short period after the ow is initiated, the local particle concentration has not yet changed enough to a ect the segregation, resulting in a relatively high segregation velocity, while at later times the particles reach a steady segregated state where the segregation velocity is negligible. We focus on the initial rapidly segregating transient. The shear rate is chosen so that the initial segregation velocity is slow enough that the 1 s sampling window used here falls within the interval of rapid segregation ( 10 s). The linear pro le of the segregation distance in Fig. 3(a) indicates the minimal e ect of the concentration change (less than 5% at most over the sampling window) on the segregation velocity. The segregation velocity of each constituent is calculated from the slope of data like that in Fig. 3(a) as w = (z    )=t, where z  is the o set at the end p;i i;1 i;1 i;1 of the t = 1 s sampling window and  is the standard error in the calculation of z  , which is represented by i;1 i;1 7 the error bars at time t slightly less than 1 s in Fig. 3(a). By considering many thin horizontal layers, the local segregation velocity pro le can be measured, as shown in Fig. 3(b) for three sample cases of uniform shear ow. The segregation velocities increase with both increasing density ratio and shear rate. For each case, segregation velocity decreases from top to bottom even though the shear rate and density ratio do not vary with depth. This is a consequence of the overburden pressure P increasing with depth, consistent with previous results [17]. Note that the sum of segregation velocities of heavy and light particles should equal zero due to volume conservation for c = c = 0:5. Figure 3(b) indeed demonstrates that the segregation h l velocities are nearly equal (and, of course, opposite). The slight di erence between w and w is likely caused p;l p;h by the small amount of segregation during the initial lling process before the start of the simulation. The error bars represent uncertainties in the measurement of the slope in Fig. 3(a). @ 2 From the momentum balance equation, Eq. (2), the convection term in the z-direction (  w ), which can be i i @z @P estimated from the simulation results, is three orders of magnitude less than the net buoyant force c   g i i i @z (noting that the constituent velocity w is equal to the segregation velocity w since the bulk velocity in the z- i p;i direction is zero). Likewise, the unsteady term (  w ) is also negligible. This con rms that the system is in i i i @t steady state during the sampling window and that the segregation velocity results from the balance between the net buoyant force and the interspecies drag as indicated by the simpli ed momentum expression in Eq. (3). A. Inertial number dependence The inertial number dependence assumed in the proposed drag model, Eq. (7), requires that the segregation velocity is linear in I [Eqs. (10) and (11)]. Therefore, to verify the drag model, segregation velocities of light particles w p;l at di erent depths are plotted versus I for di erent density ratios R in Fig. 4 for ows with nonlinear streamwise velocity pro les. Similar results are also found for the segregation velocity of heavy particles w since w  w for p;h p;h p;l ows with c = c = 0:5. The linear relation between w and I in Fig. 4 for varying _ and P (expressed in terms of h l p;i varying I ) con rms the assumed dependence on I and is consistent with results from previous studies [17, 18]. That is, the segregation velocity depends linearly on I for each density ratio. The error bars are somewhat large, consistent with the diculty in accurately measuring w as indicated in Fig. 3. Nevertheless, a line can be t through the data p;i at each density ratio. Furthermore, the tting lines are quite similar for both nonlinear velocity pro les. The data, however, do not show as clear a linear dependence on I when plotted in the same way for uniform shear 2 2 ows, as illustrated in Fig. 5(a). This is because unlike the other two nonlinear velocity pro les [u = Uz =h and 2:3(z=h1) Ue ], the local inertial number in a uniform shear ow is not equal to zero or nearly zero at the bottom wall. Instead, in the case of the linear velocity pro le (u = Uz=h), the segregation is forced to cease at the xed bottom wall because of the wall itself. To account for the cessation of density segregation at the bottom wall, we use a modi ed inertial number for uniform shear ows I = I I ; (16) where I is the inertial number at which density segregation e ectively ceases. According to Fig. 3(b), segregation ceases at z=h  0. Therefore, we assume I is the inertial number at that location, such that I = _d  =P 0 0 bulk 0 with P =  gh + P . As shown in Fig. 5(b), using I instead of I collapses the data for uniform shear ows 0 bulk wall such that the dependence of w on I is close to linear. Note that for the two nonlinear velocity pro les I equals I p;i since I  0. From here on, unless otherwise speci ed, we drop the asterisk on I , using the corrected inertial number [Eq. (16)]. Figures 4 and 5(b) show that w depends linearly on I for varying shear pro les, shear rates, and density ratios. p;i This con rms / I in Eq. (7) since equating the buoyant and drag forces gives w / I in Eqs. (10) and (11). The i p;i linear relation between w and I is also consistent with previous results in free surface ows. Consider, for example, p;i heap ows, in which the owing layer thickness is about eight particle diameters [48, 52]. For such a thin layer with the local shear rate decreasing exponentially with depth, the overburden, or lithostatic, pressure can be treated as a uniform parameter, and the linear relation between w and I can be reduced to a linear relation between w and _d p;i p;i [53, 54]. On the other hand, for a thick owing layer with a uniform shear rate, the segregation velocity at di erent depths in the owing layer is only a ected by the overburden pressure so long as we consider segregation away from the upper and lower bounding walls. The e ect of lithostatic pressure has to be included, resulting in a dependence of w on I [17, 18]. We further note that in our previous work [17], we proposed a form for I in terms of a critical p;i pressure that is equivalent to Eq. (16). However, it should be noted that at larger inertial numbers (I > 0:5) [18], instantaneous binary collisions dominate over enduring contacts, resulting in a rheological change from dense to rapid dilute granular ow. In the latter regime, the linear relation between w and I may no longer be valid. p;i 8 0 0.05 0.1 0.15 0 0.05 0.1 0.15 FIG. 4: Segregation velocities of light particles vs. local inertial number I at di erent depths for nonlinear velocity 2 2 2:3(z=h1) pro les (a) u = Uz =h and (b) u = Ue for simulations with U = 2 m/s, R =  = 2 f2; 4; 10g, h l c = c = 0:5,  = 0:2, and e = 0:9. Dashed lines are linear ts for each density ratio, demonstrating a linear h l dependence of w on I . Error bars represent the uncertainty in determining w as indicated in Fig. 3. p;l p;l 0 0.05 0.1 0.15 0 0.05 0.1 0.15 FIG. 5: Segregation velocities of light particles w at di erent depths vs. (a) inertial number I and (b) modi ed p;l 2 2 inertial number I = I I for simulations with particle properties given in Fig. 4 but for uniform shear rates 1 1 1 _ = 5 s (#), 15 s (2), and 25 s (3). I is the cut-o inertial number at which density segregation e ectively ceases [17]. The dashed line in both gures is identical and a linear t to the data in (b). Error bars represent the uncertainty in determining w as indicated in Fig. 3. p;l B. Density ratio dependence As shown in Fig. 4, w =I increases with density ratio R . Equations (10) and (11) indicate that w =I is proportional p  p to R 1=R . To test this proposed dependence, w =I , calculated as the slope of the tting lines in Figs. 4 and 5(b), p;l is plotted versus R in Fig. 6. The dashed curve in Fig. 6 represents the predictions of Eq. (10) with B( = 0:2) = 400. (We characterize the dependence of B on  in Section III.D.) Indeed, w =I is proportional to R 1=R over the p;l wide range of R tested (1:3  R  10) in di erent ow pro les where I varies due to changes in both P and _ . Good agreement between the curve and the data con rms the proposed dependence of the drag model on particle densities and, equivalently, the functional dependence of w on R in Eqs. (10) and (11). p;i  9 0 5 10 FIG. 6: w =I vs. R for simulations with 1:3  R  10,  = 0:2, e = 0:9, and c = c = 0:5. Colors and symbols p;l   h l 2 2:3(z=h1) represent simulations with _ = U=h (#), 2Uz=h (), and 2:3Ue =h (). The dashed curve shows the dependence of w =I on R from Eq. (10) with B() = 400. p;i C. Concentration ratio dependence A recent study [20] indicates that the segregation in mixtures of heavy and light particles has an underlying asymmetry that depends on the local particle concentration, similar to mixtures where particles di er in size [19]. Speci cally, a heavy particle among mostly light particles segregates faster than a light particle among mostly heavy particles. The last term  = , on the right hand side of the drag model in Eq. (7) accounts for the nonlinear h l dependence of segregation velocity on particle concentration. To con rm this term, we rewrite Eq. (10) using w = p;l c (w w ) from Eq. (5), such that c on the l.h.s and (1 c ) on the r.h.s cancel and the equation becomes h l h h l 1 gd 1 c 2 2 (w w ) = (R )I : (17) l h B()  R c solid  h According to Eq. (17), (w w ) depends linearly on c =c . To test this dependence, we keep the particle densities l h l h and other simulation parameters constant, but vary c from 0.1 to 0.9 in steps of 0.1 for uniform shear ow. Thus, c =c varies from 1=9 to 9. The particle concentration is limited to the range 0.1 to 0.9 because (w w ) is close to l h l h zero for c > 0:9, making the segregation velocity too small to accurately determine. Figure 7 plots the square of the 2 w w 2 solid 1 1 l h normalized velocity di erence between the two species w ^ = ( ) (R ) , excluding B() because I gd R , and hence B(), is the same for all simulations included in the gure, as a function of c =c for uniform shear l h ows. Note that (w w )=I is calculated using the slope of the tting line through the data at di erent depths p;l p;h like that in Fig. 5(b). This avoids propagating the random error evident in these gures as the deviation of the data points for individual simulations from the tting line. By varying only the concentration ratio in Fig. 7, it is evident p p that w ^ depends linearly on c =c . As expected, larger uncertainties exist for small c =c due to the diculty l h l h in measuring small values of the segregation velocity. Nonetheless, Fig. 7 con rms the linear relation between w ^ and c =c used in Eqs. (10) and (11), and thus the proposed form of the drag model [Eq. (7)]. l h D. Friction and restitution dependence The functional form for B() characterizing the dependence of the drag on particle friction, in Eqs. (7), (10), and (11), is dicult to derive analytically due to the complex particle interactions typical of dense granular ows. Instead, DEM simulations under a variety of ow conditions are used to determine B() and demonstrate that it is independent of the restitution coecient e. Consider rst e, which characterizes energy dissipation in short-duration collisions. We vary e from 0.1 to 0.9 for uniform shear ow while keeping all other parameters unchanged, such that the damping coecient of the linear- spring-dashpot collision model used in the DEM simulations varies accordingly. The segregation velocities for both light and heavy particles at z=h = 0:5 for the case with c = c = 0:5 and R = 8 are plotted versus e in Fig. 8(a) for h l  10 0.01 0.005 0 1 2 3 FIG. 7: Square of the normalized velocity di erence w ^ versus square root of concentration ratio c =c for l h simulations with R = 4,  = 0:2, e = 0:9, and _ = U=h = 25 s . c varies from 0.1 to 0.9 in steps of 0.1 and c = 1 c . The dashed line is a linear t to the data forced through zero. l h 0 0.5 1 0 0.3 0.6 FIG. 8: Dependence of segregation velocities w on contact parameters. (a) Light () and heavy (#) particle p;i segregation velocities are nearly independent of restitution coecient e at z=h = 0:5 with  = 0 (black), 0:1 (red), and 0:4 (blue). (b) Particle segregation velocity decreases as  increases for z=h = 0:3 (red), 0:5 (blue), 0:7 (black) with e = 0:9. R = 8, c = c = 0:5, and _ = U=h = 25 s . h l three di erent friction coecients . Although there is some scatter in the data, it is clear that w is independent of p;i e, regardless of , as is reasonable to expect for dense granular mixtures in which short-duration collisions are unlikely to play as important of a role as in dilute ows. For dense granular mixtures, enduring contacts are dominant, and stresses are generated by long-duration sliding and rolling contacts. Unlike e, the surface friction coecient  has signi cant impact on w as shown in Fig. 8(b). The p;i particle segregation velocity generally decreases as  increases at di erent bed depths. Note that w measured from p;i DEM simulations is a ected by not only  but also the  dependent initial packing, which adds more uncertainties resulting in scatter in the data. Nevertheless, the overall trend of decreasing segregation velocity with increasing friction remains evident. In this context, B() is determined through a series of DEM simulations for uniform shear ow with di erent density ratios R and friction coecients . Since the dependence of drag on the other parameters has been veri ed, it is possible to determine the functional form for B() from Eq. (10), 11 0.2 0.15 0.1 0.05 0 100 0 2 4 6 8 0 0.2 0.4 0.6 gd c 1 l 2 2 FIG. 9: (a) (R ) (1 c ) vs. (w =I ) for uniform shear ows with R 2 f2; 4; 6; 8; 10g and three l p;l R c solid  h di erent  values. B() is determined by the tted slope of the dashed lines, which are forced through zero (c = c = 0:5, _ = 25, and e = 0:9). (b) B() vs.  based on 135 simulations. Di erent symbols correspond to h l e = 0:2 (), 0.8 (), and 0.9 (#) for uniform shear ows. gd 1 c w l p;l B() = (R ) (1 c ) : (18) R c I solid  h In applying Eq. (18), the slope of the tting line in Fig. 5(b) is again used for w =I . We estimate B() by plotting p;l gd c 1 l 2 2 (R ) (1 c ) versus (w =I ) over a range of density ratios for three values of  in Fig. 9(a). The l p;l R c solid  h slope of the data at each  gives the corresponding value of B(). By measuring the slopes of di erent sets of data with 0    0:6, we empirically determine the dependence of B() on . The results are plotted in Fig. 9(b) for several values of e for uniform shear ow. For bidisperse mixtures with a small friction coecient ( < 0:3), B() is linearly proportional to . For larger values of  (0:3    0:6), the relation for B() remains linear but with a smaller slope. There is slightly more scatter for di erent values of e at larger . Note that B( = 0:2) = 400 from Fig. 9(b) is consistent with the value used in Fig. 6 to calculate w =I . p;i The interspecies drag is a combination of normal and tangential forces resulting from long-duration contacts. When = 0, the tangential forces vanish. Thus, B( = 0) 6= 0 means that interspecies drag of smooth particle systems is entirely due to normal forces. Figure 9(b) shows that B() increases with . That is, increasing friction increases the interspecies drag, thereby reducing the segregation velocity. This is contrary to observations from size segregation, where part of the segregation driving force comes from interparticle friction, such that increasing friction coecient promotes the segregation of large particles [55]. A possible explanation is that the friction force adds to the segregation driving force in size segregation, which overcomes the resultant increase in the interspecies drag. Alternately, for particles having di erent densities but the same size, the friction force has no impact on the segregation driving force but only adds to the drag. Note that Eqs. (10) and (11) are not limited to speci c velocity pro les even though B() is empirically determined using the data from uniform shear ows with c = c = 0:5. B() is merely a coecient to account for the particle h l frictional properties. Thus, Eqs. (10) and (11) are generally applicable to predict density segregation in any dense segregating granular ow, as shown in the next section. IV. SEGREGATION VELOCITY MODEL Now that the various terms of the segregation velocity model have been validated, it is possible to compare the predictions of Eqs. (10) and (11) to the segregation velocity measured in the DEM simulations. To do this, the global values for d,  , B(), c , c , and R are used in the model along with local values of I at each horizontal position solid l h  12 0.8 0.6 0.4 0.2 0 2 4 0 2 4 FIG. 10: Predictions of segregation velocity pro les from Eqs. (10) and (11) (dashed curves) compared to DEM simulation data for light () and heavy (#) particles under (a) uniform ( _ = U=h) and (b) varying shear rate 2 2:3(z=h1) pro les [ _ =2Uz=h , 2:3Ue =h] with di erent density ratios.  = 0:2, c = c = 0:5, and e = 0:9. h l 0.8 0.6 0.4 0.2 0 2 4 6 8 0 2 4 6 8 FIG. 11: Predictions of segregation velocity pro les from Eqs. (10) and (11) (dashed curves) compared to DEM simulation data under uniform shear for (a) light and (b) heavy particles with heavy particle concentration c =0.3, 0.5 and 0.7 (c = 1 c ).  = 0:2, R = 8, _ = 25 s , and e = 0:9. l h in the ow. The local inertial number I is based on the overburden pressure estimated as P = P +  g(h z), wall bulk such that I can be expressed as a function of vertical position z according to Eq. (8) (or, equivalently, Eq. (16) for uniform shear ows). Figure 10(a) compares the particle segregation velocity pro les predicted by Eqs. (10) and (11), with those measured from DEM simulations of uniform shear ows in Fig. 3(b). Since data for uniform shear ows is used in determing B(), good agreement between the predicted segregation velocities and those from uniform shear ows is expected. However, the segregation velocities for z=h < 0:2 and z=h > 0:8 in the simulations are a ected by the segregating particles accumulating near the bounding walls. 2 2:3(z=h1) A more rigorous test of the model is to consider ows with varying shear rates [ _ = 2Uz=h , 2:3Ue =h] in Fig. 10(b). The close match between the prediction and the data demonstrates that Eqs. (10) and (11) are also e ective for ow with non-uniform shear rates, even though B() in the model was derived from ow with uniform shear rate ( _ = U=h). Figure 11 provides a similar comparison where the concentration ratio c =c is varied while keeping other parameters h l constant for uniform shear. In this case, the segregation velocities for heavy and light particles di er from one another when the concentration of the two species is not equal, as expected [20]. Again, the estimated segregation velocities 13 match the DEM data. Thus, we conclude that the segregation model [Eqs. (10) and (11)] based on the interspecies drag term derived from analogy with KTGF can be used to estimate the segregation velocities of both light and heavy species through the depth of the owing layer under a variety of ow conditions. Unlike the con ned shear ows discussed to this point, the shear rate and the concentration ratio of many free surface ows vary across the domain. To test the model under these more general conditions, we consider quasi-2d bounded one-sided heap ow in which particles ow in a thin surface layer down a slope much like what occurs when lling a silo [53, 56{58]. Figure 12 shows segregation velocities from heap ow DEM simulations [52] versus 1=4 (c =c ) (1 c )I , since, according to Eqs. (10) and (11), these two variables are linearly related when all other l h i conditions are equal. The di erence between these results and those up to this point is that these results correspond to a wide range of shear rates and concentrations, all occurring simultaneously in the owing layer of the heap. The solid lines in Fig. 12 are based on the model, the data points represent DEM results [52], and the dashed lines represent least squares ts through the data points. Overall, the segregation velocities of both bounded heap ows are reasonably well predicted by the model. However, since segregation in heap ows mainly occurs in a thin layer at the free surface, the averages are more uncertain, so the data have substantial scatter, especially for the weakest segregation case with R = 1:84 and d = 3 mm in Fig. 12(a). The case with a small particle diameter d = 2 mm but large density ratio R = 3:3 has overall larger w such that the DEM results match the theoretical predictions p;i better, as shown in Fig. 12(b). Note that we only include the heap ow data for 0:1  c ; c  0:9 to match the h l concentration range for ows upon which our model is based. As Fig. 12 shows, the segregation model we propose in this paper works reasonably well not only for the con ned shear ow between two planes from which it is developed, but also for free surface ow in a bounded heap. Previous results [52] suggest that the segregation velocity can be modeled as jw j = S _ (1 c ) (19) p;i D i for bounded heap ow with inlet concentration c = c = 0:5. Equations (10) and (11) have this form when S , the l h D segregation length scale, is 1=2 S 1 1 c  gd D l bulk = R : (20) d B() R c P solid  h Previous results for free surface ows [52] indicate that the empirically determined segregation length scale S is well approximated by = C lnR ; (21) where C = 0:081. The empirical relation and the underlying data are shown in Fig. 13(a). To compare S and S it is necessary to account for the pressure and particle concentration ratio, as well as D D other ow conditions. The comparison is made by rst assuming c =c = 1, which is the feed concentration ratio in l h these simulations, such that S is no longer concentration dependent. Furthermore, in heap ows, the segregation occurs in a thin owing layer with a thickness of only a few particle diameters, in which the local shear rate decreases exponentially with depth while the scaled local pressure P=( gd) increases linearly from 1 at the surface to =d at bulk the bottom of the owing layer at depth , assuming constant  . The mean shear rate of the owing layer equals bulk the local shear rate at a depth of 0:26 due to the exponential streamwise velocity pro le. This depth corresponds to approximately 2d for 7d    8:5d [52]. By assuming the exponentially varying local shear rate is the dominant factor for the depth varying segregation velocity, we expect that the e ective pressure is equivalent to the pressure at a depth of about 2d such that P=( gd) = 2. For comparison, P=( gd) = 1 and 3 are also considered. bulk bulk The expression for S from Eq. (20) for c =c = 1 and three di erent pressures is compared to bounded heap ow D h l results [52] in Fig. 13(b). The data match the model curve well for P=( gh) = 2. Again, the results in Fig. 13(b) bulk demonstrate the success of the model in replicating a wide range of results even down to the slight curvature evident in the data matching that of the curve corresponding to Eq. (21). 14 light light -2 heavy heavy -4 0 0.04 0.08 0.12 0.16 0 0.04 0.08 0.12 0.16 FIG. 12: Segregation velocities from DEM heap ow simulations (data points) [52]. (a) R = 1:84,  = 0:2, e = 0:9, and d = 3 mm. (b) R = 3:33,  = 0:2, e = 0:9, and d = 2 mm. Dashed lines are linear ts to the data points, and solid lines are predictions of the model. 0.3 0.3 0.2 0.2 0.1 0.1 0 0 -0.1 -0.1 -0.2 -0.2 -0.3 -0.3 -1 1 -1 1 10 1 10 10 10 FIG. 13: Dependence of segregation length scale S on the density ratio R from DEM simulations (circles) for segregation in a bounded heap ow [52] compared to predictions of (a) the empirical model Eq. (21) and (b) Eq. (20) derived from the segregation velocity model with c =c = 1,  = 0:2 and di erent values of P=( gd). h l bulk V. CONCLUSION In this paper, we have proposed a predictive model, Eqs. (10) and (11), for the segregation velocities of light and heavy particle species in density bidisperse granular ows that is based on a new model for the interspecies drag in segregating dense ows, Eq. (7). The interspecies drag model assumes that the multiple long-duration particle interactions in dense granular ows re ect similar physics to short-duration binary particle interactions typical of dilute granular ows, which have been successfully modeled using the Kinetic Theory of Granular Flows (KTGF) [29, 30]. Of course, particle segregation depends on the pressure-shear state, which can be characterized by the inertial number I . In particular, I is inversely proportional to the square root of the overburden pressure, which can signi cantly reduce the segregation velocity [17]. In addition, the segregation depends non-linearly on the local particle concentration: heavy particles among many light particles segregate more quickly than light particles among many heavy particles [20]. Finally, particles in dense granular ows experience enduring contacts characterized by interparticle friction, which is in contrast to dilute granular ows where particle contacts are short and dominated by the elastic properties of the particles. Hence, the restitution coecient e has little in uence on segregation in the 15 dense ows considered here. The advantage of the proposed interspecies drag model for dense granular ows [Eq. (7)] is that it links interspecies drag to segregation velocities and includes the e ects of the local ow condition, I , local concentration ratio, c =c , h l and particle properties including the particle densities, sizes, and surface friction, . The segregation velocities derived from combining the interspecies drag model with the equilibrium momentum balance equation match the segregation velocities determined from DEM simulations in both con ned shear and free surface heap ows. This allows calculation of segregation velocities through the depth of the owing layer for di erent density ratios and relative constituent concentrations. Despite these advances, the modi ed KTGF drag model proposed here is not without drawbacks. The segregation velocities predicted from Eqs. (10) and (11), while showing the right trends, are only reasonable estimates, as is evident from the scatter in the DEM simulation data from the model predictions in Figs. 10-12. This is likely a result of the many variables in the problem, as well as the stochastic nature of the forces on individual particles that drive the segregation and the concurrent collisional di usion. Nevertheless, the results in Fig. 13 demonstrate a remarkable correspondence between the segregation velocity model and the simulation data, particularly since the data are for a di erent ow than that from which the model was derived. More research is needed under even more widely varying conditions to re ne the model. Furthermore, the model is limited to density bidisperse granular materials. A more challenging problem is to connect the interspecies drag to the segregation velocity for size-bidisperse particles, or, even more dicult, particles that di er in both size and density. [1] J. M. N. T. Gray and A. R. Thornton, A theory for particle size segregation in shallow granular free-surface ows, Proc. R. Soc. A 461, 1447 (2005). [2] S. W. Meier, R. M. Lueptow, and J. M. Ottino, A dynamical systems approach to mixing and segregation of granular materials in tumblers, Adv. Phys. 56, 757 (2007). [3] Y. Fan and K. M. Hill, Theory for shear-induced segregation of dense granular mixtures, New J. Phys. 13 (2011). [4] D. R. Tunuguntla, O. Bokhove, and A. R. Thornton, A mixture theory for size and density segregation in shallow granular free-surface ows, J. Fluid Mech. 749, 99 (2014). [5] H. Xiao, D. McDonald, Y. Fan, P. B. Umbanhowar, J. M. Ottino, and R. M. Lueptow, Controlling granular segregation using modulated ow, Powder Technol. 312, 360 (2017). [6] Y. Zhao, H. Xiao, P. B. Umbanhowar, and R. M. Lueptow, Simulation and modeling of segregating rods in quasi-2d bounded heap ow, AIChE J. 64, 1550 (2018). [7] H. Xiao, Y. Fan, K. V. Jacob, P. B. Umbanhowar, M. Kodam, J. F. Koch, and R. M. Lueptow, Continuum modeling of granular segregation during hopper discharge, Chem. Eng. Sci. 193, 188 (2019). [8] P. B. Umbanhowar, R. M. Lueptow, and J. M. Ottino, Modeling segregation in granular ows, Annu. Rev. Chem. Biomol. Eng. 10 (2019). [9] D. V. Khakhar, J. J. McCarthy, and J. M. Ottino, Radial segregation of granular mixtures in rotating cylinders, Phys. Fluids 9, 3600 (1997). [10] D. V. Khakhar, J. J. McCarthy, and J. M. Ottino, Mixing and segregation of granular materials in chute ows, Chaos 9, 594 (1999). [11] A. Tripathi and D. V. Khakhar, Density di erence-driven segregation in a dense granular ow, J. Fluid Mech. 717, 643 (2013). [12] P. Gajjar and J. M. N. T. Gray, Asymmetric ux models for particle-size segregation in granular avalanches, J. Fluid Mech. 757, 297 (2014). [13] J. M. N. T. Gray and C. Ancey, Particle-size and-density segregation in granular free-surface ows, J. Fluid Mech. 779, 622 (2015). [14] B. Marks, P. Rognon, and I. Einav, Grainsize dynamics of polydisperse granular segregation down inclined planes, J. Fluid Mech. 690, 499 (2012). [15] J. M. N. T. Gray, Particle segregation in dense granular ows, Annu. Rev. Fluid Mech. 50, 407 (2018). [16] T. Weinhart, S. Luding, and A. R. Thornton, From discrete particles to continuum elds in mixtures, in AIP conference proceedings, Vol. 1542 (AIP, 2013) pp. 1202{1205. [17] A. M. Fry, P. B. Umbanhowar, J. M. Ottino, and R. M. Lueptow, E ect of pressure on segregation in granular shear ows, Phys. Rev. E 97, 062906 (2018). [18] S. Liu and J. J. McCarthy, Transport analogy for segregation and granular rheology, Phys. Rev. E 96, 020901 (2017). [19] K. van der Vaart, P. Gajjar, G. Epely-Chauvin, N. Andreini, J. M. N. T. Gray, and C. Ancey, Underlying asymmetry within particle size segregation, Phys. Rev. Lett. 114, 238001 (2015). [20] R. P. Jones, A. B. Isner, H. Xiao, J. M. Ottino, P. B. Umbanhowar, and R. M. Lueptow, Asymmetric concentration dependence of segregation uxes in granular ows, Phys. Rev. Fluids 3, 094304 (2018). [21] D. Gidaspow, M. Syamlal, and Y. Seo, Hydrodynamics of uidization of single and binary size particles: supercomputer modeling, in Proceedings of Fluidization V (AIChE Engineering Foundation, 1986) pp. 1{8. [22] M. Syamlal, The particle-particle drag term in a multiparticle model of uidization, Tech. Rep. (DOE/MC/21353-2373:Work 16 perfomed under contract No: DE-AC21-85MC21353, 1987). [23] L. Huilin and D. Gidaspow, Hydrodynamics of binary uidization in a riser: CFD simulation using two granular temper- atures, Chem. Eng. Sci. 58, 3777 (2003). [24] L. Huilin, Z. Yunhua, J. Ding, D. Gidaspow, and L. Wei, Investigation of mixing/segregation of mixture particles in gas{solid uidized beds, Chem. Eng. Sci. 62, 301 (2007). [25] H. Iddir, H. Arastoopour, and C. M. Hrenya, Analysis of binary and ternary granular mixtures behavior using the kinetic theory approach, Powder Technol. 151, 117 (2005). [26] H. Iddir and H. Arastoopour, Modeling of multitype particle ow using the kinetic theory approach, AIChE J. 51, 1620 (2005). [27] Z. Chao, Y. Wang, J. P. Jakobsen, M. Fernandino, and H. A. Jakobsen, Investigation of the particle{particle drag in a dense binary uidized bed, Powder Technol. 224, 311 (2012). [28] J. M. N. T. Gray and V. Chugunov, Particle-size segregation and di usive remixing in shallow granular avalanches, J. Fluid Mech. 569, 365 (2006). [29] J. T. Jenkins and S. B. Savage, A theory for the rapid ow of identical, smooth, nearly elastic, spherical particles, J. Fluid Mech. 130, 187 (1983). [30] C. K. K. Lun, S. B. Savage, D. J. Je rey, and N. Chepurniy, Kinetic theories for granular ow: inelastic particles in couette ow and slightly inelastic particles in a general ow eld, J. Fluid Mech. 140, 223 (1984). [31] C. h. Liu, S. R. Nagel, D. A. Schecter, S. N. Coppersmith, S. Majumdar, O. Narayan, and T. A. Witten, Force uctuations in bead packs, Science 269, 513 (1995). [32] J. T. Jenkins, Dense shearing ows of inelastic disks, Phys. Fluids 18, 103307 (2006). [33] J. T. Jenkins, Dense inclined ows of inelastic spheres, Granul. Matter 10, 47 (2007). [34] J. T. Jenkins and D. Berzi, Dense inclined ows of inelastic spheres: tests of an extension of kinetic theory, Granul. Matter 12, 151 (2010). [35] D. Vescovi, D. Berzi, P. Richard, and N. Brodu, Plane shear ows of frictionless spheres: Kinetic theory and 3D soft-sphere discrete element method simulations, Phys. Fluids 26, 053305 (2014). [36] Y. Duan, Z.-G. Feng, E. E. Michaelides, and S. Mao, Modi ed kinetic theory applied to the shear ows of granular materials, Phys. Fluids 29, 043302 (2017). [37] Y. Duan and Z.-G. Feng, A new kinetic theory model of granular ows that incorporates particle sti ness, Phys. Fluids 31, 013301 (2019). [38] M. Larcher and J. T. Jenkins, Segregation and mixture pro les in dense, inclined ows of two types of spheres, Phys. Fluids 25, 113301 (2013). [39] M. Larcher and J. T. Jenkins, The evolution of segregation in dense inclined ows of binary mixtures of spheres, J. Fluid Mech. 782, 405 (2015). [40] M. Syamlal, W. Rogers, and T. J. OBrien, MFIX documentation: Theory guide, National Energy Technology Laboratory, Department of Energy, Technical Note DOE/METC-95/1013 and NTIS/DE95000031 (1993). [41] O. Owoyemi, L. Mazzei, and P. Lettieri, CFD modeling of binary- uidized suspensions and investigation of role of particle{ particle drag on mixing and segregation, AIChE J. 53, 1924 (2007). [42] F. Hern andez-Jim enez, L. Garc a-Guti errez, A. Soria-Verdugo, and A. Acosta-Iborra, Fully coupled TFM-DEM simulations to study the motion of fuel particles in a uidized bed, Chem. Eng. Sci. 134, 57 (2015). [43] J. L. Lebowitz, Exact solution of generalized percus-yevick equation for a mixture of hard spheres, Phys. Rev. 133, A895 (1964). [44] D. Gera, M. Syamlal, and T. J. O'Brien, Hydrodynamics of particle segregation in uidized beds, InterNatl. J. (Wash.) of Multiphase Flow 30, 419 (2004). [45] GDR MiDi, On dense granular ows, Eur. Phys. J. E. 14, 341 (2004). [46] F. da Cruz, S. Emam, M. Prochnow, J.-N. Roux, and F. Chevoir, Rheophysics of dense granular materials: Discrete simulation of plane shear ows, Phys. Rev. E 72, 021309 (2005). [47] P. Jop, Y. Forterre, and O. Pouliquen, A constitutive law for dense granular ows, Nature 441, 727 (2006). [48] Y. Fan, P. B. Umbanhowar, J. M. Ottino, and R. M. Lueptow, Kinematics of monodisperse and bidisperse granular ows in quasi-two-dimensional bounded heaps, Proc. R. Soc. A 469, 20130235 (2013). [49] L. Jing, C. Y. Kwok, Y. F. Leung, and Y. D. Sobral, E ect of geometric base roughness on size segregation, in EPJ Web of Conferences, Vol. 140 (EDP Sciences, 2017) p. 03056. [50] K. van der Vaart, A. R. Thornton, C. G. Johnson, T. Weinhart, L. Jing, P. Gajjar, J. M. N. T. Gray, and C. Ancey, Breaking size-segregation waves and mobility feedback in dense granular avalanches, Granul. Matter 20, 46 (2018). [51] L. Jing, C. Y. Kwok, Y. F. Leung, and Y. D. Sobral, Characterization of base roughness for granular chute ows, Phys. Rev. E 94, 052901 (2016). [52] H. Xiao, P. B. Umbanhowar, J. M. Ottino, and R. M. Lueptow, Modelling density segregation in owing bidisperse granular materials, Proc. R. Soc. A 472, 20150856 (2016). [53] Y. Fan, C. P. Schlick, P. B. Umbanhowar, J. M. Ottino, and R. M. Lueptow, Modelling size segregation of granular materials: the roles of segregation, advection and di usion, J. Fluid Mech. 741, 252 (2014). [54] C. P. Schlick, Y. Fan, A. B. Isner, P. B. Umbanhowar, J. M. Ottino, and R. M. Lueptow, Modeling segregation of bidisperse granular materials using physical control parameters in the quasi-2d bounded heap, AIChE J. 61, 1524 (2015). [55] L. Jing, C. Y. Kwok, and Y. F. Leung, Micromechanical origin of particle size segregation, Phys. Rev. Lett. 118, 118001 (2017). [56] Y. Fan, Y. Boukerkour, T. Blanc, P. B. Umbanhowar, J. M. Ottino, and R. M. Lueptow, Strati cation, segregation, and 17 mixing of granular materials in quasi-two-dimensional bounded heaps, Phys. Rev. E 86, 051305 (2012). [57] Z. Deng, P. B. Umbanhowar, J. M. Ottino, and R. M. Lueptow, Continuum modelling of segregating tridisperse granular chute ow, Proc. R. Soc. A 474, 20170384 (2018). [58] Z. Deng, P. B. Umbanhowar, J. M. Ottino, and R. M. Lueptow, Modeling segregation of polydisperse granular materials in developing and transient free-surface ows, AIChE J. 65, 882 (2019).

Journal

Condensed Matter – arXiv (Cornell University)

Published: Jun 8, 2019

There are no references for this article.