A cellular automaton for segregation during granular avalanches

Segregation is a complex and poorly understood phenomenon that is prevalent in many industrial and natural granular flows. When grains flow down a slope [1–5], are spun in a rotating drum [6–8] or shaken in a box [9], we observe those grains organising into intriguing patterns. Kinetic sieving is the dominant mode of segregation in granular avalanches, where separation of particles occurs according to size. Using a cellular automaton we have modelled kinetic sieving as the swapping of particles in a one-dimensional system. From the cellular automaton we have deduced a continuum model to describe the segregation.     

Discrete element modeling of non-linear submerged particle collisions

Coupling of the Discrete Element Method with Computational Fluid Dynamics (DEM–CFD) is a widely used approach for modeling particle–fluid interactions.Although DEM–CFD focuses on particle–fluid interaction, the particle–particle contact behavior is usually modeled using a simple Kelvin–Voigt contact model which may not represent realistic interactions of particles in high viscosity fluids. This paper presents an implementation of a new user-defined contact model that accounts for the effects of lubrication of fluid between two approaching particles while maintaining all other DEM–CFD particle–fluid interaction phenomena. Theoretical model that yields a non-linear restitution coefficient for submerged particle collisions, which was developed by Davis et al. (J Fluid Mech 163:479–497, 1986), is implemented in a DEM–CFD code. In this model, the behavior of particles at a contact depends on fluid properties, particle velocities and distance between particle surfaces. When two particles approach each other in a fluid, their kinetic energy decreases gradually because of a lubrication effect associated with the thin fluid layer between the particles. Particle post-collision behavior is governed by a simplified elastic contact law. With lubrication, it is possible that particles are not able to rebound if the approaching velocity is completely damped by lubrication, and in this case the particles agglomerate in the fluid. Tangential surface friction-slip forces are activated as in the case of dry particle contact. The lubrication model represents an advanced submerged particle collision approach that permits improved accuracy when modeling problems with high particle concentrations in a fluid. An application of the new model is shown in a simple sediment transport problem.     

Mathematical modeling of coating time in dry particulate coating using mild vibration field with bead media described by DEM simulation

To theoretically understand the previously reported dry particulate coating process using a mild vibration field with a bead media, a mathematical analysis model of the dry coating system was developed. In this coating process, an ordered mixture with coarse host particles (drug-loaded ion exchange resin, diameter approximately 100 µm) and fine guest particles (acrylic polymer particle, primary particle size of approximately 100 nm) is formed using a vibrating a vessel. Second, the guest particles on the host particulate surface are firmly fixed using the collision of coated particles zirconia beads (diameter 1.5 mm). Our model assumes that the unfixed guest particles are fixed by particle-to-particle collisions (C_c) provided by the apparatus, thereby increasing the coating ratio. C_c was estimated using the discrete element method and some experimental results. The model includes parameters such as the number of C_c, host particles and unfixed guest particles. The coating time simulated by the established model equation in this study fits well with the experimental results of the dry process. It depends on the ratio of the number of collisions contributing to the increased coating ratio to the number of unfixed guest particles on the surface of host particles.     

Study of powder flow patterns in a Couette cell with axial flow using tracers and solid fraction measurements

We present different aspects of dense granular flows in a Couette geometry using a variety of particulate materials with shape and size distributions. Tracer studies point to an apparent coupling of particle size with flow and stress field gradients. While there is a clear industrial motivation to use “real” materials as a means to expand basic physical and engineering research in granular dynamics, the current study suggests additional academic motivations. Indeed, particles with distributed characteristics uncover rich interactions between flow and stress fields that might otherwise go un-noticed with model materials such as spherical glass beads. Distribution of size and shape play a strong role in how stress is transmitted in granular media (Kheiripour Langroudi et al. in Powder Technol 203:23–32, 2010) and how particle pattern arrangements evolve. Direct solid fraction measurements, using a capacitance probe, show that dense particle flows exhibit significant variations in solid fraction in both sheared and stagnant layers. Furthermore, these measurements also show different dependence of the solid fraction on shearing rate: solid fraction decreases in sheared layers and increases in stagnant layers as the shear rate is increased. From these results the thickness of the shear band could be estimated and was found to vary as a function of particle shape and the roughness of the container walls. The main result is that shear stress (or torque) (see also Kheiripour Langroudi et al. in Powder Technol 197:91–101, 2010) and solid fraction profiles depend on particle shape and whether or not an extra degree of freedom in their movement is provided so that the system can dilate under various shear states in the Couette cell. This extra degree of freedom is assured in the present experimental work by allowing a slight axial outflow from the Couette device while the driven shear fields are in the radial and tangential directions.     

An implementation of the soft-sphere discrete element method in a high-performance parallel gravity tree-code

We present our implementation of the soft-sphere discrete element method (SSDEM) in the parallel gravitational N-body code pkdgrav, a well-tested simulation package that has been used to provide many successful results in the field of planetary science. The implementation of SSDEM allows for the modeling of the different contact forces between particles in granular material, such as various kinds of friction, including rolling and twisting friction, and the normal and tangential deformation of colliding particles. Such modeling is particularly important in regimes for which collisions cannot be treated as instantaneous or as occurring at a single point of contact on the particles’ surfaces, as is done in the hard-sphere discrete element method already implemented in the code. We check the validity of our soft-sphere model by reproducing successfully the dynamics of flows in a cylindrical hopper. Other tests will be performed in the future for different dynamical contexts, including the presence of external and self-gravity, as our code also includes interparticle gravitational force computations. This will then allow us to apply our tool with confidence to planetary science studies, such as those aimed at understanding the dynamics of regolith on solid celestial body surfaces, or at designing efficient sampling tools for sample-return space missions.     

ACOUSTIC FLOW INSTRUMENTS FOR SOLID/GAS FLOWS

Abstract

Two nonintrusive acoustic flow sensing techniques are reported. One technique, passive in nature, simply measures the bandpassed acoustic noise level produced by particle/particle and particle/wall collisions. The noise levels, given in true root-mean-square voltages or in autocorrelations, show a linear relationship to particle velocity but increase with solid concentration. Therefore, the passive technique requires calibration and a separate measure of solid concentration before it can be used to monitor particle velocity. The second technique is based on the active cross-correlation principle. It measures particle velocity directly by correlating flow-related signatures at two sensing stations. The velocity data obtained by this technique are compared with measurements by a radioactive-particle time-of-flight (TOF) method. A multiplier of 1.53 is required to bring the acoustic data into agreement with the radioactive TOF result. The difference may originate from the difference in flow fields where particles are detected. The radioactive method senses particles mainly in the turbulent region and essentially measures average particle velocity across the pipe, while the acoustic technique detects particles near the pipe wall, and measures the particle velocity in the viscous sublayer. Both techniques were tested in flows of limestone and air and glass beads (1 mm) and air at the Argonne National Laboratory Solid/Gas Flow Test Facility. The test matrix covered solid velocities of 20 to 30 m/s in a 2-in. pipe and solid-to-gas loading ratios of 6 to 22.     

A Lagrangian modeling approach with the direct simulation Monte-Carlo method for inter-particle collisions in turbulent flow

A Lagrangian modeling approach, which combines the direct simulation Monte-Carlo (DSMC) method and a Reynolds-averaged Navier-Stokes model to account for inter-particle collisions and turbulence characteristics of the carrier fluid, respectively, is proposed. The wall-bounded turbulent particle-laden flows in which the experimental data are available are chosen as the test problems for demonstration. Results obtained with the deterministic method accounting for inter-particle collisions are used as a basis for validating the proposed stochastic Lagrangian model. Good agreement between the predictions obtained separately with the deterministic and DSMC methods is achieved. The benefit of saving computational expenditure when using the DSMC method becomes more remarkable than the deterministic method as the number of particles loaded in the flow is increased. In addition, the study demonstrates that τ_P/τ_C is a proper parameter to monitor the role of inter-particle collisions in the physical processes of particle-laden flows.     

EXPERIMENTAL AND NUMERICAL STUDY OF THE PARTICLE FLUCTUATING MOTION INDUCED BY COLLISIONS BETWEEN PARTICLES IN A VERTICAL GAS-SOLID FLOW

ABSTRACT

The effect of interparticle collisions on the gravitational motion of large particles in a vertical convergent channel is experimentally and numerically investigated. A probabilistic collision model is implemented in a three-dimensional Monte Carlo type Lagrangian simulation code. The numerical predictions are compared to the experimental results. It is shown that an interparticle collision model is necessary to reproduce the experimentally observed particle fluctuating motion characteristics. The simulation results using the present probabilistic collision model are found to yield satisfactory agreement with experimental observations, even though the collision frequency seems to be slightly overestimated. In particular, reduction of initial anisotropy of the particle fluctuating motion with increasing particulate mass flow rate is well reproduced by the simulation. A rather good agreement is also observed between experimental results and quantitative predictions of statistical properties of the flow such as particle axial and transverse velocity distributions and standard deviations.     

Analysis of triboelectric charging of particles due to aerodynamic dispersion by a pulse of pressurised air jet

Triboelectric charging of powders causes nuisance and electrostatic discharge hazards. It is highly desirable to develop a simple method for assessing the triboelectric charging tendency of powders using a very small quantity. We explore the use of aerodynamic dispersion by a pulse of pressurised air using the disperser of Morphologi G3 as a novel application. In this device particles are dispersed by injection of a pulse of pressurised air, the dispersed particles are then analysed for size and shape analysis. The high transient air velocity inside the disperser causes collisions of sample particles with the walls, resulting in dispersion, but at the same time it could cause triboelectric charging of the particles. In this study, we analyse this process by evaluating the influence of the transient turbulent pulsed-air flow on particle impact on the walls and the resulting charge transfer. Computational Fluid Dynamics is used to calculate particle trajectory and impact velocity as a function of the inlet air pressure and particle size. Particle tracking is done using the Lagrangian approach and transient conditions. The charge transfer to particles is predicted as a function of impact velocity and number of collisions based on a charge transfer model established previously for several model particle materials. Particles experience around ten collisions at different velocities as they are dispersed and thereby acquire charges, the value of which approaches the equilibrium charge level. The number of collisions is found to be rather insensitive to particle size and pressure pulse, except for fine particles, smaller than about 30 µm. As the particle size is increased, the impact velocity decreases, but the average charge transfer per particle increases, both very rapidly. Aerodynamic dispersion by a gas pressure pulse provides an easy and quick assessment of triboelectric charging tendency of powders.     

Parallelising the MARS15 code with MPI for shielding applications

The MARS15 Monte Carlo code capabilities to deal with time-consuming deep penetration shielding problems and other computationally tough tasks in accelerator, detector and shielding applications, have been enhanced by a parallel processing option. It has been developed, implemented and tested on the Fermilab Accelerator Division Linux cluster and network of Sun workstations. The code uses a message passing interface MPI. It is scalable and demonstrates good performance. The general architecture of the code, specific uses of message passing and effects of a scheduling on the performance and fault tolerance are described.     

Remanence properties of substituted Ba-ferrite particles for highdensity magnetic recording

In this paper, the remanence properties of Co-Sn, Co-Ti and Co-Ti-Sn substituted Ba-ferrite (BaF) oriented particulate samples are compared with those of some oriented acicular particulate samples. A new parameter, the minor remanence distribution (MRD), is proposed to review the remanence properties of magnetic particles and the capabilities for resisting the recording demagnetization of magnetic recording media. It is shown that the MRD values of the oriented BaF particulate samples were smaller compared to oriented Co-γ-Fe₂O₃ samples, even though the squareness ratios (SR) of some of the BaF samples were smaller than those of the Co-γ-Fe₂O₃ samples. It Is the small MRD, SFD_r, IRS and large DH _r of a medium that can result in a large resistance to the effects of recording demagnetization and therefore in superior characteristics for high density magnetic recording. Since Co-Sn substituted BaF platelet-shaped particles exhibit these characteristics and have a very low temperature coefficient of coercivity, these particles can be expected to be a promising candidate for high density magnetic recording     

The free (open) boundary condition (FBC) in viscoelastic flow simulations

The Free (or Open) Boundary Condition (FBC, OBC) was proposed by Papanastasiou et al. (A New Outflow Boundary Condition, Int. J. Numer. Meth. Fluids, 1992; 14:587–608) to handle truncated domains with synthetic boundaries where the outflow conditions are unknown. In the present work, implementation of the FBC has been extended to viscoelastic fluids governed by explicit differential constitutive equations. As such we consider here the Criminale-Ericksen-Filbey (CEF) model, which also reduces to the Second-Order Fluid (SOF) for constant material parameters. The Finite Element Method (FEM) is used to provide numerical results in simple Poiseuille flow where analytical solutions exist for checking purposes. Then previous numerical results are checked against Newtonian highly non-isothermal flows in a 4:1 contraction. Finally, the FBC is used with the CEF fluid with data corresponding to a Boger fluid of constant material properties. Particular emphasis is based on a non-zero second normal-stress difference, which seems responsible for earlier loss of convergence. The results with the FBC are in excellent agreement with those obtained from long domains, due to the highly convective nature of viscoelastic flows, for which the FBC seems most appropriate. The FBC formulation for fixed-point (Picard-type) iterations is given in some detail, and the differences with the Newton–Raphson formulation are highlighted regarding some computational aspects.     

Computational techniques and validation of 3-dimensional viscous/turbulent codes for internal flows

Computational techniques and codes developed for the prediction of three-dimensional turbulent flows in internal configurations and rotor passages are described. Detailed calibration and validation of the flow fields in 90° curved ducts, cascades, end-wall flows and turbomachinery rotors are presented. Interpretation and comments on accuracy, level of agreement with various turbulence models and limitations of the codes are described. The single pass space-marching code is found to be efficient for curved duct and two-dimensional cascade flows. Multipass space-marching, time-marching and zonal methods are found to be accurate for complex situations. The efficiency and accuracy of a zonal technique, with considerable saving in computational time, is demonstrated. Paper presented atagard Symposium “Validation of Computational Fluid Dynamics,” May 2–5, 1988, Lisbon, Portugal The research work on computation was sponsored by the David Taylor Naval Ship Research and Development Center with Dr D Fuhs as the technical monitor, and NASA Lewis Research Center with Dr P Sockol as the technical monitor.     

Development and characterization of stabilized double loaded mPEG-PDLLA micelles for simultaneous delivery of paclitaxel and docetaxel

Objective: Double loaded micelles (DLM) in which paclitaxel (PTX) and docetaxel (DTX) were co-solubilized with monomethoxy poly(ethylene glycol)-block-poly(d,l-lactide) (mPEG-PLA) copolymer were prepared and evaluated in an aim to investigate the effect of a combination of PTX and DTX on the stability of mPEG-PLA micelles compared to single drug-loaded micelles (SDM), especially that recent clinical anticancer formulations are limited by the existence of toxic excipients and stability issues.

Materials and methods: The SDM and DLM of PTX and DTX were prepared by a solvent evaporation method. Micellar size, size distribution, drug loading content and drug release were investigated. Transmission electron microscopy was used to investigate the stabilization mechanism.

Results: The drug loading efficiency of both PTX and DTX in DLM and SDM were 25% and 10%, respectively. ¹H NMR showed a successful encapsulation of both drugs in the polymeric micelle. DLM showed better physical stability at drug concentrations higher than 1?mg/mL compared to SDM. Moreover, DLM, SDM-PTX and SDM-DTX were stable for 24, 9 and 1?h, respectively. The stabilization mechanism of DLM was investigated, a network structure of DLM was observed in TEM graphs. Furthermore, DLM showed complete and faster drug release compared to SDM. mPEG-PLA double loaded micelles can deliver two poorly water soluble anticancer drugs at clinically relevant doses. The obtained results offer a promising alternative for double drug therapy without any formulation associated undesirable effects and encourage further in vivo development and optimization of the DLM as a drug delivery system for anticancer drugs.     

A Simple Index for Measuring Particulate Contamination in Parenteral Solutions

Abstract

Concern over the recent introduction of limits for allowable particulate contamination in small volume parenterals by the United States Pharmacopeia XXI has drawn attention to the physics of systems containing small numbers of particles of varying (random) size and random identity dispersed in an aqueous mediun. In general, a size distribution of parenteral particulate characteristically consists of large numbers of small particles and relatively few particles at larger sizes. In parenteral solutions, the size distributions found tend to fit a linear power law curve of the type

In N = C - M In D

where N = cumulative number of particles/unit volume at size threshold D, M is the slope of the distribution and C is a constant. A numerical index of contamination can be obtained by integrating this curve between limits, effectively an area under the curve, AUC.

The limits are defined as the number of particles/unit volume (mL) at a size of 1.Oμ, and the size at which the count falls to 1 per mL. By simple geometry, the area of a triangle,

AUC = 1/2 [(1n N_1.0) × (1no_1.0)]

Use of this AUC is illustrated by examples taken from the literature, experimental data generated for the purpose and archived quality control data. Differences on the same materials examined by different instrumental principles may be explained by shape factors inevitably involved when analyzing samples contaminated with randon identify particulate. The AUC may provide a tool for exploring this feature of parenteral solution. Some initial experiment evaluations of the concept are provided.     

A coupled CFD-DEM analysis of granular flow impacting on a water reservoir

Massive debris flows or rock avalanches falling into a water reservoir may cause devastating hazards such as overtopping or dam breakage. This paper presents a coupled Computational Fluid Dynamics and Discrete Element Method (CFD-DEM) analysis on the impacting behaviour of a granular flow falling from an inclined slope into a water reservoir. The coupling between CFD and DEM considers such important fluid–particle interaction forces as the buoyancy force, the drag force and the virtual mass force. It is found that the presence of water in the reservoir can generally help to reduce direct impact of granular flow on the check dam behind the reservoir, minimizes the intense collisions and bouncing among particles and helps form a more homogeneous final deposited heap as compared to the dry case. While the interparticle/particle–wall frictions and collisions dominate the energy dissipation in the dry granular flow, the majority of kinetic energy of the granular system in the wet case is first transferred to the water body, which leaves the granular flow itself to become a contact-shearing dominant one and causes impulse wave travelling between the check dam and the slope surface for a rather sustained period before settling down. A power law distribution is found for the velocity profile of the granular flow travelling on both the slope and the reservoir ground surfaces, and it may change temporarily to a linear distribution at the transition point of the slope toe where the Savage number depicts a peak. The consideration of rolling friction among particles may homogeneously reduce the travelling velocity of the granular flow and alleviate the overall impact on the check dam. The impact on the check dam depends on both the initial debris releasing height and the reservoir water level. Medium water levels in the reservoir have been found to be generally safer when the initial debris height is relatively high.     

Effect of base roughness on size segregation in dry granular flows

This paper presents simulations of dry granular flows along a sloping channel using the discrete element method. The kinetic sieving and squeeze expulsion theories are utilized to study the effects of base roughness on size segregation and the underlying mechanisms. Basal friction has a significant influence on flowing regimes inside the granular body, and a larger base friction accelerates the size segregation process. The front zone of the granular body is more likely to be collision dominated with increasing base friction; as a result, the energy dissipated by frictional shearing decreases, and damping energy due to particles collisions is enhanced. Meanwhile, granular flows become much looser, and collisions between particles increase rapidly. It is shown that the differences in the kinetics among grains of mixed sizes and the mechanical effects of particle contacts can explain the mechanism of size segregation. The parameter representing the intensity of particles exchange also increases as base friction increases. The forces acting on particles are also affected by base friction. The dimensionless contact force describing the contribution of contact channel-normal stress increases as base friction increases, which indicates that a higher dispersive trend has developed inside the granular body.     

ACOUSTIC FLOW INSTRUMENTS FOR SOLID/GAS FLOWS

Two nonintrusive acoustic flow sensing techniques are reported. One technique, passive in nature, simply measures the bandpassed acoustic noise level produced by particle/particle and particle/wall collisions. The noise levels, given in true root-mean-square voltages or in autocorrelations, show a linear relationship to particle velocity but increase with solid concentration. Therefore, the passive technique requires calibration and a separate measure of solid concentration before it can be used to monitor particle velocity. The second technique is based on the active cross-correlation principle. It measures particle velocity directly by correlating flow-related signatures at two sensing stations. The velocity data obtained by this technique are compared with measurements by a radioactive-particle time-of-flight (TOF) method. A multiplier of 1.53 is required to bring the acoustic data into agreement with the radioactive TOF result. The difference may originate from the difference in flow fields where particles are detected. The radioactive method senses particles mainly in the turbulent region and essentially measures average particle velocity across the pipe, while the acoustic technique detects particles near the pipe wall, and measures the particle velocity in the viscous sublayer. Both techniques were tested in flows of limestone and air and glass beads (1 mm) and air at the Argonne National Laboratory Solid/Gas Flow Test Facility. The test matrix covered solid velocities of 20 to 30 m/s in a 2-in. pipe and solid-to-gas loading ratios of 6 to 22.     

EXPERIMENTAL AND NUMERICAL STUDY OF THE PARTICLE FLUCTUATING MOTION INDUCED BY COLLISIONS BETWEEN PARTICLES IN A VERTICAL GAS-SOLID FLOW

The effect of interparticle collisions on the gravitational motion of large particles in a vertical convergent channel is experimentally and numerically investigated. A probabilistic collision model is implemented in a three-dimensional Monte Carlo type Lagrangian simulation code. The numerical predictions are compared to the experimental results. It is shown that an interparticle collision model is necessary to reproduce the experimentally observed particle fluctuating motion characteristics. The simulation results using the present probabilistic collision model are found to yield satisfactory agreement with experimental observations, even though the collision frequency seems to be slightly overestimated. In particular, reduction of initial anisotropy of the particle fluctuating motion with increasing particulate mass flow rate is well reproduced by the simulation. A rather good agreement is also observed between experimental results and quantitative predictions of statistical properties of the flow such as particle axial and transverse velocity distributions and standard deviations.