Discrete Element Modeling Framework Research Articles

Scale‐Dependent Air‐Sea Exchange in the Polar Oceans: Floe‐Floe and Floe‐Flow Coupling in the Generation of Ice‐Ocean Boundary Layer Turbulence

AbstractSea ice is a heterogeneous, evolving mosaic of individual floes, varying in spatial scales from meters to tens of kilometers. Both the internal dynamics of the floe mosaic (floe‐floe interactions), and the evolution of floes under ocean and atmospheric forcing (floe‐flow interactions), determine the exchange of heat, momentum, and tracers between the lower atmosphere and upper ocean. Climate models do not represent either of these highly variable interactions. We use a novel, high‐resolution, discrete element modeling framework to examine ice‐ocean boundary layer (IOBL) turbulence within a domain approximately the size of a climate model grid. We show floe‐scale effects could cause a marked increase in the production of fine‐scale three‐dimensional turbulence in the IOBL relative to continuum model approaches, and provide a method of representing that turbulence using bulk parameters related to the spatial variance of the ice and ocean: the floe size distribution and the ocean kinetic energy spectrum.

Discrete element models for elasto-plastic wave propagation in nonlinear architected materials

Nonlinear architected materials are known to exhibit a plethora of dynamic phenomena that enhance our capacity to manipulate elastic waves. Since these properties stem from complex, subwavelength geometry, dynamic simulations at high resolutions are often intractable at scales of interest. Therefore, prior studies have turned to effective medium models that capture essential properties in the long-wavelength limit. One class of such models is a periodic structure composed of discrete mass, spring, and damper elements, whose constitutive relationships are computed to match behaviors observed in experiments or fine-scale simulations of representative volume elements. While models of this type have been implemented successfully in many cases, the majority of literature has been focused on recoverable deformation, possibly including linear damping. However, history-dependent effects, such as friction and plasticity, have been much less explored. In this work, we develop a discrete element modeling framework for nonlinear architected materials undergoing plastic deformation. Due to the history- and rate-dependent characteristics of plasticity, the framework generally yields a system of differential-algebraic equations whose computational cost is significantly greater than an elastic system of comparable size. We demonstrate the method using several examples from the literature and explore means to obtain phenomenological plasticity models for general geometries.

Detailed biomass fast pyrolysis kinetics integrated to computational fluid dynamic (CFD) and discrete element modeling framework: Predicting product yields at the bench-scale

Fast pyrolysis is an intricate process due to the variability and anisotropy of lignocellulosic biomass and the complicated chemistry and physics during conversion in a bubbling fluidized bed reactor (BFBR). The complexity of biomass fast pyrolysis lends itself well to computational fluid dynamics (CFD) and discrete element (DEM) analysis, which promises to reduce experimental time and its associated cost. This study investigated switchgrass fast pyrolysis simulated by computational fluid dynamics coupled with a discrete element method to track individual reacting biomass particles throughout a bench-scale BFBR reactor. We accounted for the fast pyrolysis chemistry through a comprehensive reaction scheme with secondary cracking reactions. We performed a three-step reduction for secondary cracking reactions to convert the full cracking scheme into a reduced scheme easily incorporated into our model. We assessed the impact of operational conditions on the steady-state yields of liquid bio-oil, non-condensable gases (NCG), at 550 °C over a range of fluidization numbers (2 – 6 Umf), reported as a ratio to the minimum fluidization velocity (Umf). At steady-state, the volatile bio-oil yield had a range of 49.3–50.4 wt%. Levoglucosan was the primary volatile component present with 21 wt% of the bio-oil while water was the second largest with 20 wt%. The reduction of the secondary reaction schemes did not appreciably affect the overall yields of switchgrass pyrolysis compared to the full secondary scheme.

Micromechanical description of adsorptive-capillary stress in wet fine-grained media

The aim of this paper is to investigate the effect of an adsorbed water layer on the mechanical behavior of fine-grained wet granular materials in the pendular regime with isolated capillary liquid bridges. The adsorbed water forms a thin liquid film tightly bound to a particle’s surface equilibrated by a so-called “disjoining pressure”. In a stress transmission analysis, this disjoining pressure concept is embedded in the so-called Augmented Young-Laplace equation to account for thin film interfacial interactions. Using a homogenization technique for upscaling the micro-scale physics, an adsorptive-capillary stress tensor is derived whose discrete representation reveals a new interparticle cohesive force. In the presence of adsorbed layers, it is shown that the new liquid bridge profile, as numerically solved from the Young-Laplace equation, leads to a higher cohesive interparticle force and rupture distance. The proposed adsorptive-capillary stress tensor is further implemented within a discrete element modeling framework. As such, the evolutions of microstructure, stress tensors, and shear strength are illustrated during suction-controlled triaxial simulations. Our numerical results demonstrate that adsorbed layers have a notable effect on the mechanical behavior of fine-grained materials, particularly at higher suctions.

PyBONDEM-GPU: A discrete element bonded particle Python research framework – Development and examples

Discrete element modelling (DEM) is widely used to simulate granular systems, nowadays routinely on graphical processing units. Graphics processing units (GPUs) are inherently designed for parallel computation, and recent advances in the architecture, compiler design and language development are allowing general-purpose computation to be computed on multiple GPUs. Application of DEM to bonded particle systems are much less common, with a number of open research questions remaining. This study outlines a Bonded-Particle Research DEM Framework, PyBONDEM-GPU, written in Python. This framework leverages the parallel nature of GPUs for computational speed-up and the rapid prototype flexibility of Python. Python is faster and easier to learn than classical compiled languages, making computational simulation development accessible to undergraduate and graduate engineers. PyBONDEMGPU leverages the Numba-CUDA module to compile Python syntax for execution on GPUs. The framework enables research of fibre pull-out from fibre-matrix embeddings. Bonds are simulated between all interacting particles. The performance of PyBONDEM-GPU is compared against Python CPU implementations of PyBONDEM using the Numpy and Numba-CPU Python modules. PyBONDEM-GPU was found to be 1000 times faster than the Numpy implementation and 4 times faster than the Numba-CPU implementation to resolve forces and to integrate the equations of motion.

Modelling imperfections in unreinforced masonry structures: Discrete element simulations and scale model experiments of a pavilion vault

The structural assessment of doubly-curved vaulted masonry structures, such as pavilion vaults, poses challenges specific to their high degree of indeterminacy. Two-dimensional equilibrium analysis methods may provide a lower bound of load or displacement capacity, but they do not accurately describe the three-dimensional (3D) behaviour of these structures, particularly when shear deformation (e.g. sliding) is important. Therefore, discrete element modelling (DEM) methods, which can effectively simulate 3D load re-distribution, have been used to investigate support displacement capacity and corresponding 3D collapse mechanisms. DEM analyses are usually conducted on perfect digital geometries. Meanwhile, both real structures and small-scale physical models have implicit assembly and fabrication imperfections, which may significantly alter their response. The present paper aims to investigate the influence of geometrical and mechanical imperfections by comparing DEM analyses with the results obtained from tests on a scale model. In particular, a new method to simulate imperfections within the DEM framework is proposed, and a DEM parametric analysis is compared to the measured behaviour of a 3D-printed scale model of a pavilion (or cloister) vault on spreading supports. The influence of both mechanical imperfections andgeometrical imperfections, due to element geometry deviations or imprecision of the assembly process, have been investigated. Based on these analyses, the three-dimensional behaviour of a pavilion vault subjected to horizontal displacement of the supports is described, and the variability of results due to imperfections is demonstrated. A good agreement between DEM analyses and 3D-printed scale model tests is achieved, in terms of crack patterns and mechanisms. Geometrical imperfections did change the load paths within the vault, as expected, and also influenced the displacement capacity.

Combining advanced computational and imaging techniques as a quantitative tool to estimate patellofemoral joint stress during downhill gait: A feasibility study

BackgroundThe onset and progression of patellofemoral osteoarthritis (OA) has been linked to alterations in cartilage stress—a potential precursor to pain and subsequent cartilage degradation. A lack in quantitative tools for objectively evaluating patellofemoral joint contact stress limits our understanding of pathomechanics associated with OA. Research questionCould computational modeling and biplane fluoroscopy techniques be used to discriminate in-vivo, subject-specific patellofemoral stress profiles in individuals with and without patellofemoral OA? MethodsThe current study employed a discrete element modeling framework driven by in-vivo, subject-specific kinematics during downhill gait to discriminate unique patellofemoral stress profiles in individuals with patellofemoral OA (n = 5) as compared to older individuals without OA (n = 6). All participants underwent biplane fluoroscopy kinematic tracking while walking on a declined instrumented treadmill. Subject-specific kinematics were combined with high resolution geometrical models to estimate patellofemoral joint contact stress during 0%, 25 %, 50 %, 75 % and 100 % of the loading response phase of downhill gait. ResultsIndividuals with patellofemoral OA demonstrated earlier increases in patellofemoral stress in the lateral patellofemoral compartment during loading response as compared to OA-free controls (P = 0.021). Overall, both groups exhibited increased patellofemoral contact stress early in the loading response phase of gait as compared to the end of loading response. Results from this study show increased stress profiles in individuals with patellofemoral OA, indicating increasing joint loading in early phases of gait. SignificanceThis modeling framework—combining arthrokinematics with discrete element models—can objectively estimate changes in patellofemoral joint stress, with potential applications to evaluate outcomes from various treatment programs, including surgical and non-surgical rehabilitation treatments.

Linking particle properties to layer characteristics: Discrete element modelling of cohesive fine powder spreading in additive manufacturing

Abstract Particle properties play a key role in determining the flowability of cohesive fine powders and consequently the quality of parts fabricated by additive manufacturing. However, how the characteristics of a spread powder layer are linked to particle properties remains not well understood, largely due to the limitations of available numerical models and characterisation approaches. This study thus established an efficient discrete element modelling framework to address these issues, combining GPU computing with a novel methodology for particle stiffness scaling. The validity of the DEM model was verified on both packing and spreading of cohesive fine powders. The effect of particle cohesion on the characteristics of a spread powder layer was systematically analysed for non-cohesive and weakly to strongly cohesive powders. The structure of spread powder layer was qualitatively illustrated and quantitatively evaluated by not only commonly used metrics, such as coordination number, packing density and surface profiles but also particle clustering and pore characteristics. While reducing particle cohesion leads to an enhanced powder flowability, different regimes were identified in the relationship between layer characteristics and particle cohesion. The results showed that powder spreadability has a complicated dependence on the strength of particle cohesion, where the underlying mechanisms can be understood in part via a dimensionless inertial length. These findings not only provide valuable metrics to quantitatively evaluate the quality of a spread powder layer, but also enable a better understanding of the physics underlying the spreading of cohesive fine powders.

Numerical simulation of rock mass blasting using particle flow code and particle expansion loading algorithm

Numerical simulation of limestone rock mass blasting was carried out using the two-dimensional particle flow code (PFC2D) within the discrete element modeling framework, and the explosion load was loaded using the particle expansion algorithm. A field experiment of rock blasting was also carried out from which the size of the blasting funnel and the maximum blasting vibration velocity were recorded. The simulation results were compared with the experimental results in two aspects: the blasting result and the post-explosion effect. From the comparison, it has been observed that the size of the blasting funnel obtained from the numerical simulation shows a good agreement with the field experiment. However, in the case of the post-explosion effect (the blasting vibration velocity), the results of the numerical simulation and the vibration velocity curve obtained from the field test show disagreement to a certain extent. This demonstrates that the particle expansion loading algorithm is reliable in simulating the rock blasting process. The results of this research serve as a reference for subsequent applications of the PFC method in the field of rock blasting.

Discrete element modeling for continuous powder feeding operation: Calibration and system analysis.

Research emphases on extensive experimental studies and modeling efforts have been on the rise for the development of accurate predictive models of pharmaceutical unit operations and 'digital-twin' framework for continuous manufacturing lines. These exhaustive studies have been conducted at different process conditions to acquire comprehensive knowledge of effects of process parameters on the overall process dynamics. However, there still lacks a detailed understanding of material property effects of pharmaceutical powders on process operation. To address this issue, a discrete element modeling (DEM) approach combined with material calibration is applied for simulation of feeder unit to obtain particle-level insight into effects of material properties on feeder performance with focus on particle flow and powder mixing within the feeder unit. Bulk calibration is implemented to accurately represent powder material properties within the DEM framework. Different refill situations are simulated using DEM to observe powder mixing, measured at the outlet. Feeder DEM simulations are further applied to understand correlations of material properties on feeder operation. These studies provide a detailed physical insight and particle-scale information into the powder mechanics during powder feeding operation.

Particle-shape induced radial segregation in rotating cylinders

In this work we assess whether particle shape can induce radial segregation in a rotating cylinder. To this end, mixtures of spheres and non-spherical particles of equal-volume were modelled in a discrete element modelling framework. We could observe that particle-shape (alone) can induce radial segregation with the non-spherical particles accumulating in the centre of the cylinder. To probe the underlying segregation mechanism, a large number of particle trajectories and orientations were analysed. We observed that non-spherical particles, when segregating towards the centre of the bed, did not necessarily orientate themselves such that their projected area (in the sinking direction) was smaller than that of spheres. Instead, we found that there was a high probability that the avalanching, non-spherical particles are orientated such that their projected area in the direction perpendicular to the bed surface is maximal. Hence, particle-shape induced segregation cannot be explained by the conventional percolation mechanism. Instead, we propose that shape-induced segregation originates from the lower mobility of the non-spherical particles than spheres. We explain the lower mobility of non-spherical particles by their larger radius of gyration than equal-volume spheres, leading to an increased number of collisions and thus a higher rate of kinetic energy dissipation.

A hybrid discrete bubble-lattice Boltzmann–discrete element model for gas-charged sediments

This paper presents a hybrid discrete bubble-lattice Boltzmann–discrete element modelling framework for simulating gas-charged sediments, especially in the seabed. A discrete bubble model proposed in chemical engineering is adapted in the coupled discrete element/lattice Boltzmann method to model the migration of gas bubbles in saturated sediments involving interactions between gas bubbles and fluid/solid phases. Surface tension is introduced into the discrete bubble model in this work, so that it can handle the complex gas–fluid–solid interface. The lattice Boltzmann and discrete element methods are, respectively, employed to simulate fluid flows and mechanical behaviours of sediments. A velocity interpolation-based immerse boundary method is utilised to resolve the coupling between the fluid flow and the solid/gas phase. The proposed technique is preliminarily validated using simulations of bubble migration in fluids, which is followed by high-resolution investigations of the transport of a gas bubble in seabed sediments. It is demonstrated that this hybrid method can reproduce, to a certain degree, the characters of bubbles moving in seabed sediment tests.

A regime map for the normal surface impact of wet and dry agglomerates

The normal surface impacts of wet and dry agglomerates are simulated in a discrete element modeling framework. While the impact behavior of dry agglomerates has been addressed previously, similar studies on wet agglomerate impact are missing. By adding a small amount of liquid to a dry agglomerate, the impact behavior changes significantly. The impact behavior of the agglomerates at different moisture contents and impact energies are analyzed through postimpact parameters and coupled to their microscopic and macroscopic properties. While increasing the impact energy breaks more interparticle bonds and intensifies damage and fragmentation, increasing the moisture content is found to provide the agglomerates with higher deformability and resistance against breakage. It is shown that the interplay of the two latter parameters together with the agglomerate structural strength creates various impact scenarios, which are classified into different regimes and addressed with a regime map. © 2018 American Institute of Chemical Engineers AIChE J, 64: 1975–1985, 2018

Towards stochastic discrete element modelling of spherical particles with surface roughness: A normal interaction law

The current work is the first attempt towards establishing a stochastic discrete element modelling framework by developing a normal contact interaction law based on the classic Greenwood and Williamson (GW) model for spheres with rough surfaces. Two non-dimensional forms of the model that have a substantial impact on the computational efficiency are discussed and the theoretical relationship between the GW model and the Hertzian model for smooth spheres is formally established. Due to the inter-dependence between the contact pressure and deformation distributions in the model, a Newton–Raphson based iterative solution procedure is proposed to effectively and accurately obtain the contact force in terms of the overlap and two surface roughness parameters. The related key components of the procedure are addressed in detail. The numerical results obtained are first validated and then curve-fitted to derive an empirical formula as a new normal interaction law for spheres with surface roughness. The explicit nature of the new interaction law makes it readily be incorporated into the current discrete element modelling framework. A simple example is presented to illustrate the effect of surface roughness on the packing behaviour of a particle assembly.

A method for representing boundaries in discrete element modelling—part I: Geometry and contact detection

AbstractThe discrete element method for analysis of the dynamic behaviour of discontinuous media is well established. However, its application to engineering problems is still limited to simplified representations of structural boundaries and their kinematics. In this paper a method is developed for representing three‐dimensional boundaries of arbitrary geometry and for modelling the interaction between boundary objects and particles within the discrete element modelling framework. The approach, which we term the finite wall method, uses planar triangular elements to approximate the boundary surface topology. Any number of wall elements can be used to model the shape of the structure. A contact detection scheme is presented for boundary surfaces and spheres based on a series of vector projections to reduce the problem dimensionally. The algorithm employs spatial sporting to obtain the set of potential contacts between spheres and wall elements prior to contact resolution. In a further stage, all possible contact conditions including contact with surfaces, edges and corners are explicitly determined. Part I of this two‐part series of papers describes the finite wall method for representation of surface geometry and fully elaborates the method for detecting and resolving contact between boundary wall elements and spheres. In Part II the finite wall method is extended to apply kinematics to linearly independent boundary objects using combinations of translational and rotational motion. An approach is developed for coupling the DEM with the FEM for the purpose of optimising the design of structures which are dynamically interacting with particulate media. Copyright © 2001 John Wiley & Sons, Ltd.