Particle-wall Interactions Research Articles

Adaptive AI-based surrogate modelling via transfer learning for DEM simulation of multi-component segregation.

Segregation of granular materials is a critical challenge in many industries, often aimed at being controlled or minimised. The discrete element method (DEM) offers valuable insights into this phenomenon. However, calibrating DEM models is a crucial, albeit time-consuming, step. Recently, using machine learning (ML)-based surrogate models (SMs) in the calibration process has emerged as a promising solution. Nevertheless, developing such SMs is challenging due to the high number of DEM simulations required for training. Additionally, choosing a suitable ML model is not trivial. This study aims to develop SMs that effectively link particle-particle and particle-wall DEM interaction parameters to segregation of a multi-component mixture. We evaluate several ML models, ranging from artificial neural networks to ensemble learning, that are trained on a very cost-effective dataset, employing Bayesian optimisation with cross-validation to tune their hyperparameters. Next, we introduce a novel transfer learning (TL)-based approach that leverages knowledge from a few scenarios to handle new "unseen" ones. This method enables the construction of adaptive SMs for unseen scenarios, such as a new initial configuration (IC) of granular mixtures, without the need for a full-sized dataset. Our findings indicate that Gaussian process regression (GPR) efficiently builds accurate SMs on a very small dataset. We also demonstrate that only a few samples are required to build an accurate SM for the unseen IC, which significantly reduces the data preparation burden. By incorporating one and five samples from unseen scenarios to update the TL-GPR-based surrogate model, the SM's performance (based on [Formula: see text]) on unseen scenarios improves by 17 and 47%, respectively. The insights and methodology presented in this study will facilitate and accelerate the development of accurate SMs for DEM calibration, assisting in developing reliable DEM models in a shorter timeframe.

Modeling Stochastical Particle Rebound based on High Velocity Experiments

Abstract Solid particle erosion is a major deterioration process which contributes to performance deterioration of modern axial compressors. The prediction of this deterioration process requires the correct computation of particle movement through the machine and their resulting impacts on the effected components. Especially for particles with high Stokes numbers the movement is determined mainly by the particle wall interaction, which is described by coefficients of restitution. Today, they are derived from experiments featuring high particle velocities and target materials, which are representative for turbomachinery applications. In this study, an already published rebound model is optimized for particle materials and velocities within high pressure compressors. The statistical spread of the rebound experiment is evaluated and the implementation into the rebound model is shown, which improves the prediction capability of the model. The model is implemented into a CFD software and numerical simulations are performed. The model is applied to a cylinder test specimen within a sand blast facility. The simulation shows the importance of the stochastics of the rebound, which is often neglected in particle wall models. Moreover, the numerical study shows requirements for the test specimen and its positioning in the experimental setup, which are prerequisites for the derivation of the coefficients of restitution using two dimensional particle evaluation equipment.

The use of ensembles trees machine learning method in estimating damping of granular materials

Granular Materials (GM) employed within the mechanism of particle dampers attenuate the vibration energy due to their interparticle and particle-wall interactions. Estimating their damping effect using the analytical equivalent single mass approach overlooked the particles' individual losses that are built into the total damping. Alternatively, the numerical techniques (e.g., Discrete Element) are time-inefficient and computationally demanding. Therefore, this study explores the implementation of Machine learning (ML) algorithms to estimate the damping effect of GM. The ML model in this study will rely on a Data-Driven Modeling approach (DDM) incorporating the ensemble tress nonlinear regressor method. The models' training and testing data were obtained from an experimental setup of an acrylic beam internally integrated with Stainless Steel (SS) and glass spheres undergoing low excitation amplitude (RMS <1N). The main aim was to map between seven input features (e.g. filling ratio) and one targeted output: the beam's damped frequency response. Three ensemble trees' algorithms were used to create the DDM; Decision Tree, Random Forest, and XGBoost. The hyperparameter combination based on the Gridsearch CV function increases the prediction accuracy of each model. The developed ML models provided high accuracy (86-93%) in predicting the damping effect of the granular materials spheres.

A parallel coupling framework for DEM-MBD: Model verification and application

Bulk material handling equipment can be numerically studied using the DEM-MBD method, where the particulate dynamics are solved by DEM (Discrete Element Method) and the motion characteristics of the mechanical components are modeled by MBD (Multi-Body Dynamics). In this work, a two-way coupling framework is proposed for the co-simulation of our self-developed DEM solver (DEMSLab) with the commercial MBD software (RecurDyn) using a parallel staggered coupling approach. To verify the proposed coupling method, the simulation result of a torsional spring damper system is compared with the corresponding analytical solution. After that, to assess the feasibility of this method in industrial-scale applications, a series of simulations involving a wheel loader is carried out. Different simulation conditions are adopted to reveal their impact on computational efficiency and simulation accuracy. These conditions include the mesh representation of the equipment geometry, the MBD time step size, and the coupling interval. An improved mesh simplification algorithm is proposed and used to reasonably simplify the surface mesh of the wheel loader, thus lowering the computational intensity of particle-wall interactions in DEM. The comparison results indicate that the computational time can be largely saved while the simulation accuracy remains considerably high. The MBD step size and coupling interval are crucial for numerical stability and should be selected carefully.

Numerical study of sheath formation in multi‐ion species plasmas

AbstractA study of multi‐ion species plasmas in divertor region through kinetic simulation helps us understand particle transports and wall interactions. We analyzed plasma sheath behavior without collisions involving electrons, hydrogen isotopes, and helium ions using a one‐dimensional spatial space and three‐dimensional velocity space (1D3V) Particle‐In‐Cell (PIC) simulation. The PIC simulation model follows Maxwellian velocity distributions with the pre‐sheath acceleration for each particle species in the plasma source, and the plasmas move to the absorption wall with equal and constant flux. This revealed spatial potential variations due to differences in masses and charges of multi‐ion species plasmas, including independent sound velocities of each ion species. Increasing ion masses result in a more negative wall potential. The electrostatic force repels electrons and accelerates multi‐ions to reach the absorption wall. This information is found in the phase spaces of velocity in the sheath.

Neural network architecture search model for thermal radiation in dense particulate systems

In the CFD-DEM simulation of the many dense particulate systems, particle-scale thermal radiation is an important heat transfer mode under high temperatures. In this work, neural network architecture search model with different layer connection matrix is applied to the view factor regression of the thermal radiation in dense particulate systems. Deep neural networks with 3346 feasible architectures are evaluated by the HyperBand algorithm to find the local optimal solution. Neural architecture search model trained by the big data of the view factor gives a good prediction of the macroscopic radiative properties and it is in general agreement with the empirical correlations and experimental data. The particle–wall radiation decreases strongly with the distance and the maximum interaction depth is about 2.0 times the sphere diameter. The trained deep neural network model provides an efficient data-driven closure to discuss the thermal radiation of the particle–particle and particle–wall interactions in particle bed.

A Comparative Study of Different CFD Codes for Fluidized Beds

Fluidized beds are pivotal in the process industry and chemical engineering, with Computational Fluid Dynamics (CFD) playing a crucial role in their design and optimization. Challenges in CFD modeling stem from the scarcity or inconsistency of experimental data for validation, along with the uncertainties introduced by numerous parameters and assumptions across different CFD codes. This study navigates these complexities by comparing simulation results from the open-source MFIX and OpenFOAM, and the commercial ANSYS FLUENT, against experimental data. Utilizing a Eulerian–Eulerian framework and the kinetic theory of granular flow (KTGF), the investigation focuses on solid-phase properties through the classical drag laws of Gidaspow and Syamlal–O’Brien across varied parameters. Findings indicate that ANSYS Fluent, MFiX, and OpenFOAM can achieve reasonable agreement with experimental benchmarks, each showcasing distinct strengths and weaknesses. The study also emphasizes that both the Syamlal–O’Brien and Gidaspow drag models exhibit reasonable agreement with experimental benchmarks across the examined CFD codes, suggesting a moderated sensitivity to the choice of drag model. Moreover, analyses were also carried out for 2D and 3D simulations, revealing that the dimensional approach impacts the predictive accuracy to a certain extent, with both models adapting well to the complexities of each simulation environment. The study highlights the significant influence of restitution coefficients on bed expansion due to their effect on particle–particle collisions, with a value of 0.9 deemed optimal for balancing simulation accuracy and computational efficiency. Conversely, the specularity coefficient, impacting particle–wall interactions, exhibits a more subtle effect on bed dynamics. This finding emphasizes the critical role of carefully choosing these coefficients to effectively simulate the nuanced behaviors of fluidized beds.

Induced-charge electrophoresis of a tilted metal nanowire near an insulating wall.

Electric fields are commonly used to control the orientation and motion of microscopic metal particles in aqueous suspensions. For example, metallodielectric Janus spheres are propelled by the induced-charge electro-osmotic flow occurring on their metallic side, the most common case in electrokinetics of exploiting symmetry breaking of surface properties for achieving net particle motion. In this work, we demonstrate that a homogeneous metal rod can translate parallel to a dielectric wall as a result of the hydrodynamic wall-particle interaction arising from the induced-charge electro-osmosis on the rod surface. The applied electric field could be either dc or low-frequency ac. The only requirement for a nonvanishing particle velocity is that the axis of the rod be inclined with respect to the wall, i.e., it cannot be neither parallel nor perpendicular. We show numerical results of the rod velocity as a function of rod orientation and distance to the wall. The maximum particle velocity is found for an orientation of between ∼30^{∘} and ∼50^{∘}, depending on the position and aspect ratio of the cylinder. Particle velocities of up to tens of µm/s are predicted for typical conditions in electrokinetic experiments.

Spatially resolved TALIF investigation of atomic oxygen in the effluent of a CO2 microwave discharge

A diagnostic setup for one-dimensionally spatially resolved two-photon absorption laser-induced fluorescence (TALIF) detection of ground state oxygen atoms ( 2p43P2,1,0 ) is developed. The goal of this study is to investigate the evolution of temperatures and absolute number densities of oxygen atoms along the effluent of a low-pressure CO2 microwave discharge in order to gain insights into some of the mechanisms governing the post-discharge regime. The plasma source is operated at conditions of 600 W– 1200 W of absorbed power with flow rates of 74 sccm and 370 sccm pure CO2 at pressures between 1.2 mbar and 5 mbar with specific energy inputs up to 111.9 eV/molecule. These operating conditions exhibit high CO2 conversions (up to 90%) at low energy efficiencies (2%–7.4%), due to direct electron impact dissociation driving the conversion process resulting in splitting of CO2 into CO and metastable oxygen atoms. The TALIF measurements yield spatially resolved translational temperatures between 1000 K– 1600 K for most operating conditions and axial positions along the effluent. Reference measurements with xenon 6p′[3/2]2 are used for absolute number density calibration. The resulting axially resolved number density profiles of ground state atomic oxygen increase along the effluent, even at considerable distances of several centimeters from the active discharge, before they reach a maximum between 5×1020 m−3 and 2.2×1021 m−3 depending on the condition, and decrease after that. This behavior indicates the potential significance of quenching of metastable oxygen atoms within the post-discharge regime of the investigated CO2 discharges. The measured spatially resolved number density evolutions are qualitatively consistent with quenching via wall collisions being the dominant deactivation mechanism, underlining the importance of particle-wall interactions.

Study on Wear Characteristics of a Guide Vane Centrifugal Pump Based on CFD–DEM

Guide vane submersible centrifugal pumps are a kind of submersible pump, and the fluid inside the pump is often mixed with gravel and other impurities during operation, affecting the pump’s operating efficiency and life expectancy. However, past studies on solid–liquid two-phase flow (STF) and wear characteristics in guided vane centrifugal pumps have been limited to the particle trajectory and wear region distribution. These studies have lacked research on the effect of particles on the fluid flow and the specific amount of wear on the overflow components. Additionally, most of them have used the DPM discrete-phase model, which does not consider the particle–particle and particle–wall interactions. This paper is based on the CFD–DEM method, combined with the Archard wear model. A solid–liquid two-phase flow simulation is carried out for pumps with different particle sizes and particle shapes to analyze the particle movement inside the pump, the wear distribution and average wear amount of the overflow components, and the effect of particles on the turbulent kinetic energy of the fluid. The results show that the particles mainly collide with the leading and trailing edges of the impeller blades and the leading edge of the guide vane blades and form a buildup at the trailing edge of the concave surface of the guide vane blades, resulting in the wear being mainly distributed in these regions. With an increase in particle size and a decrease in sphericity, the average wear on the overflow components increases. The change of particle size directly affects the resistance of the fluid and the structure of the flow field, which has a large impact on the fluid flow pattern and generates large turbulent kinetic energy fluctuations. The shape of the particles only changes the structure of the local flow field, which has a small impact on the fluid flow pattern.

Experimental study of the effect of particle–wall interactions on inertial particle dynamics in wall turbulence

Based on Voronoi analysis, the properties related to the near-wall motion of particles in a turbulent boundary layer were experimentally investigated via different release modes, with a friction Reynolds number $Re_\tau =3530$ . For high-inertia sand particles with Stokes number $St^+ \sim O(10^2\unicode{x2013}10^3)$ and a volume fraction $\varPhi _v \sim O(10^{-4})$ , particle image tracking velocimetry was used to determine the particle position and near-wall distribution properties. We established three particle release modes, including top-released, bottom overall-released and bottom partially released sand particles, under the same flow field conditions and calculated the differences in particle near-wall clustering and void properties. It was confirmed that wall effects (including collision and strike-splash) have a great influence on particle clustering and void behaviour near the wall. In the top-released sand particle and locally laid sand particle cases, particles bounced off the smooth walls and re-entered the carrier flow, causing significant clustering and sparsing of particles near the walls. In contrast, in the overall sand-laying case where the bottom wall was completely covered with sand particles, there is no apparent cluster or void phenomenon near the wall $(z/\delta <0.12)$ and the particles are randomly distributed, due to the combined effect of particle impact and splashing. In addition, the clustering and voids of particles become more pronounced with increasing wall-normal distance in the three release modes, and the particle distribution shows some self-similarity at each flow layer. The probability density function of the concentration of cluster particles decreases following a ‘ $-5/3$ ’ power law. However, due to the particle–wall interaction, the probability density function gradually deviates from the ‘ $-5/3$ ’ power law.

Particle removal from solid surfaces via an impinging gas jet pulse: Comparison between experimental and CFD-DEM modeling results

Computational and experimental studies are performed to investigate the influence of a gas jet pulse impinging perpendicularly onto a flat solid surface seeded with a monolayer of spherical particles. A numerical resuspension model is derived to predict particle release into the gas phase during jet impingement. To this end, discrete particles are immersed in a continuous gas phase modeled via Large Eddy Simulation to capture the effect of jet turbulence on the particle detachment process. Two-way coupling between particulate and gas phase is established via a near-wall particle drag model and a lift model. Direct particle–wall interactions are captured with adhesion, rolling friction and sliding friction models. To validate the overall modeling approach, numerical results are compared with resuspension experiments for monodisperse polystyrene particles placed on a glass slide. Simulations show that the particles preferentially mobilize by rolling, followed to a limited extend by lift off from the solid surface driven by aerodynamic forces and particle–particle collisions. Resuspension occurs in the first instants after jet impingement. Computational and experimental results for removal efficiency Γ are in good agreement in terms of the location of their r50 parameter, i.e., the radial position where 50% of the particles have been removed by the jet; both results show a linear dependence between the r50 value and the jet Reynolds number of the impinging jet. However, experimental Γ(r) curves generally have a smooth sigmoidal shape whereas numerical results predict a sharp transition. Some model shortcomings are identified which lead to an underprediction of particle lift-off and which explain the observed differences. Furthermore, the experimental Γ curves nearly collapse onto each other when plotted over the predicted local wall-shear stress.

A coupled DEM-CFD analysis of asphaltene particles agglomeration in turbulent pipe flow

We performed numerical investigations of the agglomeration behaviour of asphaltene particles in a turbulent pipe flow by using a commercial discrete element modelling (DEM) software, EDEM, coupled with a computational fluid dynamics (CFD) software package, FLUENT, in four-way coupling. We describe the particle–particle and particle–wall interactions to the Hertz–Mindlin approach extended with the Johnson–Kendall–Roberts contact force model. The very small size of asphaltene particles, however, requires minor time steps resulting in excessively long computing times. To reduce the latter, we developed a novel scaling method in which (a) the value of the shear modulus is reduced while the non-dimensional Cohesion number is kept constant by also reducing the value of the interfacial energy, and (b) the values of the friction parameters are adapted. We calibrated our scaling method by means of two tests, viz. the stop distance of a single particle sliding and rolling over a flat surface, and a common angle of repose test, both for particles 100 and 60 µm in size. As a result, we were able to increase both the Rayleigh time step and the DEM time step by a factor of 25. The eventual DEM/CFD simulations for 5 µm particles in a turbulent pipe flow still took almost 4 months on a 48 cores workstation. They resulted in visualisations and in quantitative data on, among other things, number and size (gyration radius) of agglomerates, their fractal dimension, and a coordination number, all as a function of time.

Characterization and modeling of plasma sheath in 2.45 GHz hydrogen ECR ion sources

In this article, we present a multi-fluid numerical model developed for application on electron cyclotron resonance ion sources (ECRIS). The 1D-model is matured to compute the density of the ion species in the plasma sheath in the presence of an inhomogeneous magnetic field of a 2.45 GHz ECRIS. The multi-fluid model in cylindrical coordinates is focused on solving the continuity and momentum equations of hydrogen plasma particles to characterize their sheath properties. In addition, 28 important processes, including volume and surface collisions, have been included in the COMSOL Multiphysics package to simulate the ECR plasma. We study the elementary processes containing electron–atom, electron–molecule, atom–molecule, molecule–molecule, and particle–wall interactions. Then, the results of the model and the simulation of a 2D-hydrogen plasma are reported, and future perspectives are discussed throughout the paper.

The modulation of coherent structures by the near-wall motions of particles

Particle–wall interaction generates strong particle near-wall motion, including collision bounce and impact splashing. To distinguish the effect of particles and particle near-wall motions on the turbulent coherent structure, this study carried out three different cases of sand-laden two-phase flow measurements: a uniform sand release at the top, local-laying sand bed and global-laying sand bed (Liu et al., J. Fluid Mech., vol. 943, 2022, A8). Based on large field of view particle image velocimetry/particle tracking velocimetry measurements, we obtained the velocity field of a two-dimensional gas–solid two-phase dilute faction flow $(\varPhi _{v} \sim O(10^{-4}))$ with a friction Reynolds number $R e_{\tau }$ of 3950. Results indicate that particles weaken the high- and low-velocity iso-momentum zones and hairpin vortices, resulting in the increased length scale of the coherent structure. However, the collision bounce and impact splashing break up the inner iso-momentum zone and hairpin vortices while enhancing them in the outer region, thus reducing the structure scale. In addition, the upward-moving particles increase the large-scale structure inclination angle, while the downward-moving particles decrease it. The linear coherence spectrum analysis suggests that the particles themselves do not change the structural self-similarity, but their saltation motions disrupt the similarity of the near-wall structure, making the inclination angle decrease with the scale, and the generated ascending particles reduce the aspect ratio of the streamwise to wall-normal direction in the outer region.

Evaluating the influence of filter membrane on dust particle deposition and detachment based on CFD-DEM method

The motion of dust particles significantly affects the service life and filtration efficiency of filter elements. By employing a coupled computational approach combining Computational Fluid Dynamics and Discrete Element Method (CFD-DEM), the Johnson-Kendall-Roberts (JKR) model was incorporated into the calculation of adhesive forces between particle-particle and particle-wall interactions. The influence of the membrane layer on dust particle deposition and detachment on the surface of filter elements was investigated through simulation analysis. The influence of membrane layers on the deposition and detachment of dust particles in filter elements was investigated. The results show that the addition of the membrane layer led to a 35.5% increase in the initial filtration pressure drop of the filter. However, it concurrently resulted in a diminution of the deposition depth of dust particles. By the eighth cycling period, the deposition depth of the filter element with membrane (FWM) was 2.5 mm, which was only half that of the filter element without membrane (FOM). Moreover, it was found that the cleaning efficiency of the FOM is more sensitive to reverse flow pressure, while the cleaning efficiency of the FWM is more sensitive to the duration of pulse width. These findings can help explain the mechanism of dust particle deposition and detachment and provide a basis for the design and manufacturing of filter elements.

Particle-wall alignment interaction and active Brownian diffusion through narrow channels.

We numerically examine the impacts of particle-wall alignment interactions on active species diffusion through a structureless narrow two-dimensional channel. We consider particle-wall interaction to depend on the self-propulsion velocity direction whereby some specific particle's alignments with respect to the boundary walls are stabilized more. Further, the alignment interaction is meaningful as long as particles are close to the confining boundaries. Unbiased diffusion of active particles for various possible stable velocity alignments against the walls has been examined. We show that for the most stable configuration leading to the self-propulsion velocity direction perpendicular to the wall, diffusivity becomes inversely proportional to the square of the alignment interaction torque. On the other hand, when the self-propulsion velocity direction making an acute angle to the channel walls is the most stable configuration, diffusion exponentially grows with strengthening alignment interaction. Hence, particle-wall interaction plays a pivotal role in the transport control of active particles through narrow channels. Moreover, the impacts of the alignment interactions on diffusion largely depend on the particle's self-propulsion properties and its chirality. Our simulation results can potentially be used to understand unbiased diffusion of artificial or living micro/nano-objects (such as virus, bacteria, Janus particles, etc.) though narrow confined structures.

The influence of frequency and gravity on the orientation of active metallo-dielectric Janus particles translating under a uniform applied alternating-current electric field.

Theoretical and numerical models of active Janus particles commonly assume that the metallo-dielectric interface is parallel to the driving applied electric field. However, our experimental observations indicate that the equilibrium angle of orientation of electrokinetically driven Janus particles varies as a function of the frequency and voltage of the applied electric field. Here, we quantify the variation of the orientation with respect to the electric field and demonstrate that the equilibrium position represents the interplay between gravitational, electrostatic and electrohydrodynamic torques. The latter two categories are functions of the applied field (frequency, voltage) as well as the height of the particle above the substrate. Maximum departure from the alignment with the electric field occurs at low frequencies characteristic of induced-charge electrophoresis and at low voltages where gravity dominates the electrostatic and electrohydrodynamic torques. The departure of the interface from alignment with the electric field is shown to decrease particle mobility through comparison of freely suspended Janus particles subject only to electrical forcing and magnetized Janus particles in which magnetic torque is used to align the interface with the electric field. Consideration of the role of gravitational torque and particle-wall interactions could account for some discrepancies between theory, numerics and experiment in active matter systems.

Controlling wall-particle interactions with activity.

We theoretically determine the effective forces on hard disks near walls embedded inside active nematic liquid crystals. When the disks are sufficiently close to the wall and the flows are sufficiently slow, we can obtain exact expressions for the effective forces. We find these forces and the dynamics of disks near the wall depend both on the properties of the active nematic and on the anchoring conditions on the disks and the wall. Our results show that the presence of active stresses attract planar anchored disks to walls if the activity is extensile, and repel them if contractile. For normal anchored disks the reverse is true; they are attracted in contractile systems, and repelled in extensile ones. By choosing the activity and anchoring, these effects may be helpful in controlling the self assembly of active nematic colloids.

Effects of flow distributor structures and particle-wall interaction on baghouse gas-solid flow

The uniform distribution of gas–solid two-phase flow inside the baghouse is a key factor in its performance. The Discrete Phase Model (DPM) was applied to study the flow characteristics and uniform degree of the flow distribution. The effect of flow distributor structures, characteristics of particles (such as size and restitution coefficient), and the particle–wall interaction on gas–solid two-phase flow were investigated by experiments and simulations. Results show that flow distributors greatly improve the uniformity of fluid distribution, while showing little effect on the filtration resistance. The flow distributors effectively avoid the occurrence of jet phenomena within the baghouse and significantly reduce the quality of dust on the surface of the filter bag. The weight has decreased by approximately 24.6 % in Case B (a flow distributor with four baffles perpendicular to the flow direction) compared to Case A (without a flow distributor and baffle). More particles will deposit on the surface of the filter bag with the increasing of restitution coefficient. The particle–wall interaction has a significant impact on particle motion when the particle sizes are larger than 150 μm, while the impact on particle motion is small when the particle sizes are smaller than 10 μm. These studies provide a basis for improving the uniformity of gas–solid two-phase flow inside the baghouse.