Search

The impact of agglomerates on solid surfaces occurs inevitably during the processing of granular materials. In this work, numerical simulations of the surface impact of wet agglomerates are carried out by introducing liquid bridge interactions into a soft-sphere DEM model, and the instantaneous impact behaviors are discussed quantitatively at the single particle scale. Results indicate that the impact process usually includes four stages, i.e., compression, expansion, fragmentation, and re-agglomeration. The expanding radius is found to follow an exponential-law with time, and the number of fragments follows a power-law with impact velocity. Due to liquid bridge interactions, chain-like structures are observed beyond the impact, similar to those formed during the impact of dense suspension droplets. Moreover, the impact behaviors can be generally classified into four groups with respect to impact velocity, i.e., minor compression without splashing, severe compression with minor splashing, severe splashing with minor fragmentation, and severe fragmentation.

The clogging mechanism has attracted increasing attention due to widespread applications of fibrous filtration in the industry. We carry out computational fluid dynamics with a discrete element method to investigate the clogging in a two-fiber system. The role of interparticle adhesion among microparticles in the unclogging-clogging transition is examined. A clogging phase diagram in the form of the Stokes number and the adhesion parameter (Ad) is built. We show that the formation of clogs contains two sequential processes, the growth of particle chain and the formation of bridges connecting two fibers. A stronger adhesion results in a longer growing time but a shorter bridging time of particle chains. These competing trends yield an “optimal” Ad range with the fewest penetrating particles. The structural analysis of the clogs shows that as Ad increases, the average local volume fraction and coordination number decrease. The reconstruction of particles mainly happens in the initial cake filtration stage.

Inertial deposition of wet microparticles on a fiber is numerically investigated by means of discrete element method, taking the effects of impact velocity and initial spin into consideration. Results indicate that the overall, normal and tangential coefficients of restitution remain zero as the impact velocity is lower than a critical value, beyond which these restitution coefficients will increase nonlinearly with increasing the impact velocity. As the impact angle increases, the normal critical impact velocity will show a parabolic decrease and the tangential critical impact velocity will increase linearly, different from dry collisions where the critical impact velocities almost keep constant. As the impact velocity increases, the capture efficiency of wet microparticles by the fiber decreases exponentially, meanwhile the deposit microstructures tend to be denser. It becomes much more complex if clockwise or anti-clockwise initial spins are encountered. The rotational coefficient of restitution will first decrease with increasing the rotational rate until reaching a minimum value, beyond which the rotational restitution coefficient will increase as the rotational rate increases further. However, the overall restitution coefficient changes slightly if the rotational rate is less than about 105 rad/s. The effect of clockwise spin on capture efficiency is found to be distinctly different from that of anti-clockwise spin, and it can be reasonably interpreted by their effects on the overall coefficient of restitution.

The effects of hydrodynamic interactions on random adhesive loose packings of monodisperse spherical micro-particles are investigated via a two-way coupled CFD-DEM approach. It is found out that the packing fraction and coordination number do not follow the regression of the adhesion number, which was proposed recently to well characterize the bulk properties of adhesive packings. Based on the balance between particle adhesion and fluid drag force, we propose a modified fluid adhesion number, Adf, which successfully normalizes the packing fraction and coordination number. Besides, the pressure drop through the packed bed shows a critical behavior, which validates the jamming threshold of fine powders, but occurs at a fixed packing fraction of ϕ ≈ 0.22 under the present protocol. The normalized permeability of the packed bed for different parameters is analyzed, which is in good agreement with the derivation of Kozeny-Carman equation.

Discrete Element Model (DEM), embedded into computational fluid dynamics (CFD), is developed to simulate the behaviors of wet granular-fluid flows in a thin spouted bed possessing two-dimensional characteristics. A cohesive contact model in presence of pendular liquid (polydimethylsiloxane oil) bridges between particles is improved on the basis of lubricated sliding friction and restitution coefficient of the wet grains. Special algorithm and array structure are designed to search the neighbors of wet grains. Then, the comparison of frequency, granular patterns, connecting networks, particle velocity profiles and spout geometry between the DEM simulations and our previous experiments are performed. The DEM simulation results indicate that the minimum spouting velocity (Ums) is almost the same for both wet and dry system, which agrees well with the prediction by a simplified formula. The voidages and gas velocities along the spout axis of wet granular system are found to be larger than those of dry one. It well explains the larger vertical particle velocities of wet grains over those of dry ones. Finally, the granular temperature distributions are discussed. High granular temperature occurs not only along the spout, but also at the fountain-annulus corner. Comparing granular temperature at different bed levels, we found that granular temperature of wet case is larger than that of dry one in most regions of spout and fountain, but it is nearly the same in annulus.

The ash fouling during coal combustion is initialized by a deposition of ‘sticky’ inner layer composed of fine particulate matter (PM). In this work, we present a quantitative investigation on the built-up of the initial deposit layer and its roles in the capture of coarse fly ash particles. The fly ash and ash deposit are sampled in a 25 kW self-sustained pulverized coal combustor for various lignite samples including Zhundong and Hami lignites. The ash collection efficiency onto the probe is positively correlated to the ultrafine particle formation fraction. The deposition rate of fine particle PM10 is modeled in a combined Eulerian-Lagrangian scheme, with an input particle concentration measured at the entrance of boundary layer. The predicted results divulge that the thermophoresis enhances the impaction efficiencies of submicron and micron sized particles by 1–2 orders. The deposited particles are more concentrated on the central top region of the probe, and the trend is even more remarkable for larger particles, consistent with experimental observations. For both Zhundong and Hami lignites, the deposits are largely composed of fine PM10 and condensed matter during the initial stage of one minute, but the weight percentages of fine particle deposits over the total deposit mass decrease significantly in the subsequent 5–10 min. Consequently, the average sticking efficiency of the coarse mode PM10 + increases asymptotically as the deposited PM10 grows to ~ 1.5 g/m2. Zhundong lignite shows higher saturated sticking efficiency of PM10 + than Hami lignite, which is possibly attributed to the different surface properties of the bulk ash particles.

With a novel 3D discrete-element method specially developed with adhesive contact mechanics, random loose packings of uniform spherical micron-sized particles are fully investigated. The results show that large velocity, large size or weak adhesion can produce a relatively dense packing when other parameters are fixed, and these combined effects can be characterized by a dimensionless adhesion parameter ( $Ad=\omega/2\rho_pU^2_0R$). Four regimes are identified based on the value of $Ad$: RCP regime with $Ad20$. Force distribution of these adhesive loose packings follows $P(f)\sim f^\theta$ for small forces and $P(f)\sim \exp^{-\beta f}$ for big forces, respectively, which shares a similar form with that in packings without adhesion but results in distinct exponents of $\theta=0.879$, $\beta=0.839$. A local mechanical equilibrium analysis shows that adhesion enhances both sliding and rolling resistance so that fewer neighbours are needed to satisfy the force and torque balance., Comment: 12 pages, 10 figures

The collision-sticking phenomenon among micron-sized particles is ubiquitous in nature and plays an important role in formation of particle agglomerates or deposits. In this paper, a validated 3D JKR-based discrete-element method (DEM) is employed to investigate oblique collisions of micron-sized particles. Special attention is paid to the effects of impact parameter and particle size on the sticking/rebound criterion and energy dissipation pathways. Various energy dissipation mechanisms, including viscoelastic effect, sliding resistance and rolling resistance, are incorporated in our DEM model. Based on our simulation results, the temporal evolution for the collision process is revealed in detail to establish a deeper understanding of collision dynamics. Moreover, two regimes are clarified according to the effect of impact angles on the sticking criterion. In the low impact angle regime, the normal critical velocity V CN keeps as a constant. While in the regime of high impact angle, V CN rapidly drops. Furthermore, the strong effect of particle size ratio on the critical velocity is also analyzed. This effect can be well described by the dimensionless adhesion parameter, Ad . Finally, a generalized formula is drawn as the sticking/rebound criterion for collisions of fine particles, which can be readily implemented in computational fluid dynamics (CFD) codes.

Nanoporous TiO 2 thin films are deposited directly onto substrates by a one-step stagnation flame synthesis with organometallic precursors. The deposition mechanism in the stagnation-point boundary layer is intensively studied. For the first time, the radial profile of nanoparticle deposition flux is measured using a novel method of concentric collecting rings, which exhibits similar trend with the heat flux profile of stagnation-point flows. Then, we develop the mathematical model of nanoparticle transport and deposition in the stagnation-point boundary layer for further clarifying experimental results, especially the effects of substrate temperatures and in-situ produced particle sizes. Both thermophoresis in an inner part of boundary layer and thermal compression/expansion of the gas phase are found to play important roles in determining the deposition flux. The contribution of Brownian diffusion, determined by a thermophoretic Peclet number, is inappreciable compared to thermophoresis until particle diameter is as small as 2 nm. The results in this work support a conclusion of size-independence of the thermophoretic velocity, implying that the rigid-body collision assumption of Waldmann's formula is not accurate for small particles especially less than 10 nm. This study can be generally applied to other deposition techniques of thin films.

A JKR-based dynamic model, with energy dissipation through both irreversible snap-on/snap-off process and viscoelastic effect, is developed to simulate the dynamics of low-velocity normal impact of micro-sized particle with a flat surface. The first-contact energy loss because of the incompletely reversible, quasi-static loading/unloading is incorporated in the dissipation model. The damping forces corresponding to both the attractive and repulsive components of the JKR contact forces are introduced to account for the hysteresis related to material viscoelasticity. The predicted particle behaviors for the impacts with incident velocities below or above the critical velocity are discussed. The predictions of both the critical velocity and the variation of restitution coefficient vs. incident velocity are compared with experiments in literature where reasonable agreement is obtained. The dependence of restitution coefficient on incident velocity is analyzed. Finally, the model sensitivity analysis such as the effect of the damping coefficient variation is discussed. The work provides a solid basis for the development of discrete-element-method approach of micro-sized particulate system.

A Discrete Element Method (DEM) is used together with the continuum model of turbulent fluids to simulate the periodic spouting of granular solids in a two-dimensional spouted bed. The bed is contained in a rectangular column of 152 mm width and 15 mm depth with a tapered base. Glass beads with a diameter of 2 mm are used as bed material. Simulations using the DEM together with a low Reynolds number k – e turbulence model for the fluid phase yield predictions of the unstable spout regime, characterized as a periodic upward-moving particle jet. The simulation results compare well to experimental data obtained using a particle image velocimetry (PIV) technique, including fluid flow fields, time-averaged particle velocity profiles, and spout shape. Finally, DEM predictions for distribution of drag and net force on the particles, particle concentration fields, gas velocity and turbulence field are discussed.

Particle flow behaviors in a two-dimensional spouted bed (2DSB) with draft plates were studied using both the particle image velocimetry (PIV) and the combined technique of discrete element method and fluid dynamic computation (DEM-CFD) while considering the gas turbulence effect. The bed consisted of a rectangular column, 152 mm wide and 15 mm deep, a conical section with an included 60° angle and two draft plates with a distance of 15 mm. Images of particle flow were recorded by a high speed CCD camera and analyzed using a self-developed PIV algorithm to obtain a time-averaged particle velocity field. Experiments predict that the addition of draft plates not only makes the streamline of particles in the annulus steeper, but the velocity magnitude is made smaller as well. DEM results predict well the longitudinal profile of the particle vertical velocity along the bed centerline, especially during the rapid acceleration stage at the lower part of the spout. Finally, the distributions of drag forces and net forces are introduced in this paper to explain the particle velocity profiles by PIV measurement.

Four kinds of solid particles were captured by a single fiber. The particles included two kinds of fly ash particles and two kinds of ceramic particles under three different types of charging pretreatment. The single fiber was fixed across a square cross-section glass tube. A standard continuous aerosol generator was used to disperse particles to generate a uniform aerosol. The aerosol particles from the generator were tribocharged, polarized or charged, and then passed across the fiber. A microscope and a CCD camera were used to observe the capturing process and the shape distribution of dendrites. The results showed that the deposited particles developed in different ways. In the tribocharged case, dendrite formation can be classified into three distinct stages. In the prepolarized case, straight chains were formed at a uniform spacing interval, and had high binding intensity to support very long chains. Even though the long chains fell over they still had high capturing efficiency. In the precharged case, some straight chains with some branches were formed, but their binding intensity was low and they easily fell over and broke. The results also showed how the particle diameter and shape influenced the formation of dendrites and chains.

A novel particulate trajectory model (PTM) is developed to predict axial transport and dispersion of municipal solid wastes (MSW), based on the vector analysis on particle's gravity-induced axial displacement in a single excursion. Three parts of work are extended with respect to this PTM. First, the simplified formulas about mean residence time (MRT) and material volumetric flow (MVF) are derived by incorporating statistic-averaged analysis on all repeated excursions of solids within kiln into PTM. The correctional factors—et for MRT and ef for MVF—are introduced to improve the model's validity under such practical cases, i.e. irregular MSW existence or internal-structure presence. Reasonable agreement is obtained between the empirical formulas and experiments with correlation factor in excess of 90% for all runs. Second, a stochastic PTM is extended to predict the residence time distribution (RTD) curves of segregated MSW by considering the probability of the rolling distance of individual particle. As for MSW, the main cause of axial dispersion is the segregation of rolling distance of solids, due to variation of MSW components, shapes and sizes. Finally, the optimization model for geometry design of a laboratory-scale rotary kiln pyrolyser of MSW is presented and the corresponding optimum solutions are provided.

Experiments on the influences of operational variables on the axial transport of both heterogeneous municipal solid waste (MSW) and homogenous sand are conducted in a continuous lab-scale rotary kiln cold simulator. Compared with sand, the residence time of MSW has a relatively large discrepancy with the ideal normal distribution due to the trajectory segregation of MSW components. The residence time at different axial zone is quite different due to the varied bed depth profile along the kiln length. MSW has a longer mean residence time (MRT) and a lower material volumetric flow (MVF) than sand because of the higher θd than sand. The increment of both rotating speed and kiln slope reduces MRT, and increases MVF. Exit dam has a significant impact on the MRT and the influence of internal structure group consisting of various axial ribs and circular ribs is mainly determined by the height of circular ribs. Inside wall roughness also has effect on MRT through changing the bed regimes. For a case with the certain inlet and exit bed depths, the product of MRT and MVF holds at a constant within the limits of experimental errors in spite of the changing experimental variables.