Simulating granular flows in industrial systems using the discrete element method (DEM) is computationally expensive, creating particular challenges for large scale real-time applications and design optimisation. We present a neural network surrogate model for DEM simulations based on continuous convolutions that explicitly targets applications in complex industrial device geometries including moving boundaries and variable operational speeds. The model achieves significant computational speedup over DEM while maintaining high fidelity in mass discharge predictions, mixing degree accuracy, and granular temperature predictions. Validation across hoppers, rotating drums, and cylindrical mixers demonstrates that a single model trained on combined geometries often outperforms geometry-specific models, particularly for extrapolations beyond training conditions. Inclusion of boundary velocities as an explicit input in to the surrogate model is shown to assist in preventing unphysical particle behaviour and enable robust generalisation to unseen operating conditions, addressing key limitations of previous surrogate approaches for industrial granular systems.

Multiple collisions of granules on a substrate are encountered in a wide range of applications. In this work, such multiple collisions are analysed using Collisional Smooth Particle Hydrodynamics (CSPH) to understand how deformation caused by an impact influences the collision dynamics of subsequent impacts. It is found that the collision dynamics depends on the location of the impact and the deformation of the substrate caused by the preceding impacts. The predictions of three theoretical models are also compared with CSPH to assess the accuracy of the assumptions made by the models. The theoretical model predictions are only found to be useful when the granule repeatedly impacts the same location. Since these models do not simulate the shape change during the substrate deformation they fail to accurately model the cases where multiple impacts occur at different locations.

Granular materials are crucial components in a broad variety of industrial and natural processes. However, despite their widespread importance, predicting their complex flow behaviour under different conditions remains extremely challenging. The Discrete Element Method (DEM) is the primary computational technique used to simulate granular flows, and while it can produce highly accurate predictions, it is also inherently computationally expensive. We look to overcome this computational bottleneck through the use of a neural network surrogate model for 3D granular flow simulation, and consider the industrially important use case of flow in grain hoppers. We investigate our model performance across a range of different time scales, quantifying its accuracy and generalizability to different hopper geometries. The use of deep learning techniques for the prediction of granular flow dynamics offers an excellent opportunity for providing step change increases in computational efficiency for industrial decision-making and potential application in real-time decision making in diverse manufacturing settings.

We consider the problem of design of granular materials for optimal gripping behaviour in soft robotic devices. This represents a highly challenging multi-dimensional design problem, requiring consideration of material, geometry and control strategies. Our system consists of a bag gripper composed of a flexible balloon attached to a vacuum pump, filled with granular material of non-spherical particles. At atmospheric pressure the bag is soft, flexible and deforms around a target object. When the vacuum is then applied, the granular material within the gripper densifies, freezing into a rigid shape and gripping the target object. The morphology of the particles within the granular material play a crucial role in determining the dynamics of the gripping behaviour and the overall performance of the gripper in terms of its interaction with the target object and the strength of the grip (quantified as the pull off force). In this work we consider both a real world physical gripper and a corresponding computational model (based on the Discrete Element Method), and employ NSGA-III to optimise the design of both the complex structure of the granular material within the gripper and also an optimal gripping strategy to achieve maximum gripping performance for multiple fitness criteria.

In powder bed fusion (PBF), the global packing density is an important property used to quantify the bulk behaviour of the powder. However, packing density can be highly non-uniform across the PBF system. In some cases global packing density is even ill-defined, such as in the case of spread powder layers which are typically only 1-2 particles thick and packed across a complex surface. However, an accurate calculation of the local packing density down to the resolution of individual particles can allow us to go beyond bulk descriptions and spatially quantify local powder packing density variations. In this paper, we present Set Voronoi tessellation as a precise method for calculating the local packing fraction of non-spherical particles across arbitrary boundaries. Using the Discrete Element Method (DEM) for a calibrated Ti-6Al-4 V powder model, we study the local packing variation in three critical sections of the PBF system. First, we analyse a discharging Hall flowmeter, a common apparatus used to benchmark the flowability of PBF feedstock. Second, we analyse the powder spreading process to understand how particle densities and velocities influence deposition. Lastly, we analyse the local packing variation across a realistic AM powder layer to demonstrate how layers can be digitally qualified to inform subsequent laser melting. Our work provides a novel technique to study the variations in the local packing structure of powders in dynamic PBF systems and to understand the mechanisms through which key process parameters influence final part quality.

We test grip strength and shock absorption properties of various granular material in granular jamming robotic components. The granular material comprises a range of natural, manufactured, and 3D printed material encompassing a wide range of shapes, sizes, and Shore hardness. Two main experiments are considered, both representing compelling use cases for granular jamming in soft robotics. The first experiment measures grip strength (retention force measured in Newtons) when we fill a latex balloon with the chosen grain type and use it as a granular jamming gripper to pick up a range of test objects. The second experiment measures shock absorption properties recorded by an Inertial Measurement Unit which is suspended in an envelope of granular material and dropped from a set height. Our results highlight a range of shape, size and softness effects, including that grain deformability is a key determinant of grip strength, and interestingly, that larger grain sizes in 3D printed grains create better shock absorbing materials. The data set is publicly available at https://doi.org/10.25919/tgck-2r85.

We present a new framework for learning novel operational strategies and dynamically controlling the layering process in metal additive manufacturing. Metal additive manufacturing technologies such as powder bed fusion (PBF) are generally constrained by a fixed action powder spreading process. At every layer, the print platform is lowered by a fixed amount, and the same recoating action is performed. Ideally this would lead to consistent layering and identical properties each time, but frequently process variability disrupts this procedure, leading to inconsistent layers. This can be mitigated by intelligently controlling the powder spreading process, which we achieve via a shift to digital methodologies that can reveal new process strategies and dynamically update the printer commands. We employ Bayesian optimisation as a method to build and train surrogate models for real-time control. We then demonstrate the utility of this Smart Recoating approach within an integrated simulation framework driven by realistic Discrete Element Method powder spreading simulations. Our results inform new strategies for controlling the recoater and print stage displacements, and demonstrate the potential of a digital twin control system to mitigate process variation and achieve consistent print quality in each layer.

Granular jamming has recently become popular in soft robotics with widespread applications including industrial gripping, surgical robotics and haptics. Previous work has investigated the use of various techniques that exploit the nature of granular physics to improve jamming performance, however this is generally underrepresented in the literature compared to its potential impact. We present the first research that exploits vibration-based fluidization actively (e.g., during a grip) to elicit bespoke performance from granular jamming grippers. We augment a conventional universal gripper with a computer-controllled audio exciter, which is attached to the gripper via a 3D printed mount, and build an automated test rig to allow large-scale data collection to explore the effects of active vibration. We show that vibration in soft jamming grippers can improve holding strength. In a series of studies, we show that frequency and amplitude of the waveforms are key determinants to performance, and that jamming performance is also dependent on temporal properties of the induced waveform. We hope to encourage further study focused on active vibrational control of jamming in soft robotics to improve performance and increase diversity of potential applications.

Optimising the quality of metal parts produced by additive manufacturing requires an understanding of how the powder characteristics impact both layer spreading and the subsequent powder melting. Here we present a simulation study which couples models of powder spreading (by the Discrete Element Method) and powder bed fusion by laser beam (by CFD). Previous simulations of these processes have mostly assumed idealised smooth surfaces, spherical particles and a single laser track. Here we seek a more realistic description by spreading over a previously-melted (rough) surface and by melting a number of tracks laid side-by-side to mimic the crosshatched scans typically used during part production. The effects of powder morphology (via a range of particle shapes including spheres, disks, ellipsoids and cuboids) and layer depth on spreading and subsequent melting have been investigated. We find that the fraction of laser energy transferred to the part and the melt pool volume both increase rapidly with the volume of powder deposited, due to the combined effects of multiple laser reflections and the insulating nature of the powder layer. In contrast, particle shape has very little effect on the overall melting behaviour of the powder, with the volume deposited and uniformity of coverage being the key determining factors. These observations suggest that the use of cheaper non-spherical powders is feasible provided that a sufficiently uniform coverage can be achieved at small layer depths.

The Collisional-SPH method models the combined evolution of deformation and frictional contact forces in elastoplastic granular collisions. However, there are many applications in which the granules are not spherical. In this paper, CSPH is extended to model an impact of an ellipsoidal granule on a flat deformable substrate. The CSPH spring-stiffness formulation is developed to account for the contact area eccentricity. Validation for different granule orientations and aspect-ratios are presented. The combined effect of substrate deformation and non-spherical granule shape on the contact-zone mechanics is investigated for different granule orientations and aspect ratios. It is found that the granule's orientation and aspect ratio influences the interdependence between substrate deformation and contact-force distribution and should be accounted for in collision models. (c) 2023 The Society of Powder Technology Japan. Published by Elsevier BV and The Society of Powder Technology Japan. All rights reserved.

Digital twins present a conceptual framework for product life-cycle monitoring and control using a simulated replica of the physical system. Since their emergence, they have garnered particular attention as a shift away from costly physical testing and towards the use of high fidelity simulations, sensor data and intelligent control. Metal additive manufacturing (AM), a 3D printing technology prone to defects, requires a digital twin capable of tackling issues of printed part qualification, certification and optimisation. In this paper, we evaluate the key features specific to metal AM and review the current literature of modelling, sensing, control and machine intelligence. We find that the body of research toward the development of an metal additive manufacturing (AM) digital twin can be organised logically into a hierarchy of four levels of increasing complexity. The elements composing each level require deep integration and we highlight the key enabling technologies: surrogate modelling, in-situ sensing, hardware control systems and intelligent control policies. Our proposed digital twin hierarchy for AM provides a developer framework for engineering digital twins, both for AM and other intelligent manufacturing systems. • Evaluated the research towards the development of a digital twin for metal additive manufacturing. • Literature organised into a hierarchy of four levels. • Digital Twin framework proposed for each level of the hierarchy. • Discuss the challenges, areas of impact, and future work.

Granular systems react to changes in external pressure by adapting their density through complex grain contact interactions. Granular packings subjected to small pressures are loose and fluidic, but are jammed into compressed, rigid packings at higher pressures. Common soft robotic jamming grippers are composed of a vacuum pump connected to a flexible membrane filled with granular material, e.g. ground coffee. The membrane encompasses an object, and the pump activates. The grains jam, deforming the membrane, and grasping the object, with grain morphology playing a critical role in determining gripper performance. Bespoke grippers can be designed to effectively grasp specific objects by evolving their constituent grains. Evolved grippers have to-date used exclusively monodisperse granular materials (grains with identical size and shape). However, while not conceived through evolution, polydisperse grippers comprised of natural grains varying in size and morphology can often perform better in real-world grasping experiments. We employ the Discrete Element Method and NSGA-III to optimise grasps on disparate objects by evolving distributions of superellipsoidal grains (varying both their shapes and volumes) within the gripper. Results elucidate the successful application of multi-objective evolution to design bespoke polydisperse jamming grippers, and how variations in grain surface curvatures and volumes influences grasping performance.

Elastoplastic frictional collisions are encountered in a wide range of applications. For modelling these, inclusion of both the elastoplastic material response and history-dependent collisional forces is essential. The Collisional-SPH method, originally developed for elastic collisions, incorporates frictional collision forces into SPH and accounts for their history-dependence. In this paper, CSPH is extended to model elastoplastic collisions by incorporating an elastoplastic material model. A thorough validation of this elastoplastic CSPH method is presented. CSPH is then used to analyse the influence of deformation on history-dependent collision forces. It is found that the elastoplastic deformation and the resulting shape change of the substrate alter the contact-zone mechanics by influencing the local tangential force distribution. These effects become more pronounced with increasing deformation.

Powder recoating is a key step in metal Additive Manufacturing (AM) processes where powder is spread across laser processed surfaces to add material for the next layer. Achieving the desired thin powder layers that are both sufficiently dense and uniform is essential for maintaining the requisite geometric tolerances and final part quality. In this study, we focus on the influence of the substrate surface topography by comparing spreading performance over a set of realistic surfaces. We simulate powder spreading over these surfaces using a calibrated non-spherical particle Discrete Element Model for Ti-6Al-4V that incorporates cohesion and Coulomb friction interactions between particles and surfaces. We identify the four key length scales of the recoating process determined by frictional contacts, powder size distribution, the layer thickness and the melted surface topography. We find that realistic AM surfaces show markedly different powder coverage compared to an idealised flat-plane. Rougher surfaces are found to be recoated with larger amounts of powder than smoother surfaces, as smaller particles get trapped by the grooves and valleys across the surface. Counterintuitively, we find that a finer more cohesive powder can achieve the best layer coverage over realistic surfaces - indicating that powder flowability is an incomplete measure of powder spreading performance on realistic AM surfaces. We also demonstrate how the recoating process can significantly size segregate the feedstock powder, favouring deposition of smaller sized particles on the melted surfaces.

The classical paradigm of the 'big magma tank' chambers in which the melt differentiates, is replenished, and occasionally feeds the overlying volcanos has recently been challenged on various grounds. An alternative school of thought is that such large, long-lived and largely molten magma chambers are transient to non-existent in Earth's history. Our study of stratiform chromitites in the Bushveld Complex-the largest magmatic body in the Earth's continental crust-tells, however, a different story. Several chromitites in this complex occur as layers up to 2 m in thickness and more than 400 kms in lateral extent, implying that chromitite-forming events were chamber-wide phenomena. Field relations and microtextural data, specifically the relationship of 3D coordination number, porosity and grain size, indicate that the chromitites grew as a 3D framework of touching chromite grains directly at the chamber floor from a basaltic melt saturated in chromite only. Mass-balance estimates imply that a few km thick column of this melt is required to form each of these chromitite layers. Therefore, an enormous volume of melt appears to have been involved in the generation of all the Bushveld chromitite layers, with half of this melt being expelled from the magma chamber. We suggest that the existence of thick and laterally extensive chromitite layers in the Bushveld and other layered intrusions supports the classical paradigm of big, albeit rare, 'magma tank' chambers.

The application of granular jamming in soft robotics is a recent and promising new technology offer exciting possibilities for creating higher performance robotic devices. Granular jamming is achieved via the application of a vacuum pressure inside a membrane containing particulate matter, and is particularly interesting from a design perspective, as a myriad of design parameters can potentially be exploited to induce a diverse variety of useful behaviours. To date, the effect of variables such as grain shape and size, as well as membrane material, have been studied as a means of inducing bespoke gripping performance, however the other main contributing factor, membrane morphology, has not been studied due to its particular complexities in both accurate modelling and fabrication. This research presents the first study that optimises membrane morphology for granular jamming grippers, combining multi-material 3D printing and an evolutionary algorithm to search through a varied morphology design space in materio. Entire generations are printed in a single run and gripper retention force is tested and used as a fitness measure. Our approach is relatively scalable, circumvents the need for modelling, and guarantees the real-world performance of the grippers considered. Results show that membrane morphology is a key determinant of gripper performance. Common high performance designs are seen to optimise all three of the main identified mechanisms by which granular grippers generate grip force, are significantly different from a standard gripper morphology, and generalise well across a range of test objects.

Landslides may be triggered by a variety of external factors and can lead to dramatic human and economic consequences. Systematically investigating all possible parameters that could potentially influence the extent of a landslide is costly and generally not practical. In this paper, a methodology to numerically assess the impact of potential landslides is proposed. In this approach the landslide simulations were carried out using finite element model estimates of potential failure volume as an input into a discrete element model which is used to assess run-out and consequences. A statistical design of experiments was implemented to determine the most important factors influencing the landslide extent thereby significantly reducing the number of simulations required. The method is illustrated on a case study of potential failure of an overburden dump at a coal mine. Over the parameter ranges considered, the particle–particle (rock block to rock block) friction coefficient and the volume of failing material were determined to be the most important factors influencing the extent of the landslide. A systematic analysis of varying the particle–particle friction coefficient was then undertaken to better understand the dynamics of the collapse and different types of collapses were identified between low and high inter-particle friction coefficients. The methodology proposed in this paper should be of interest to practitioners as a way to thoroughly yet efficiently identify and assess the main factors influencing a potential landslide on sites at risk.

Accurate modeling of rebound kinematics in particle-substrate collisions is essential in a wide range of applications like milling and mixing. To accurately model these real-world collisions the forces arising due to elastoplastic deformation and friction need to be modelled accurately. In this study the use of frictional boundary conditions in modeling elastic collisions with Smooth Particle Hydrodynamics (SPH) is explored. The collision dynamics for an oblique impact of a 3D spherical granule on an elastic substrate is assessed for SPH (a) without any special treatment for friction and (b) using Coulomb's friction model, and it is identified that these approaches are inaccurate in some instances. To resolve this, spring-based contact models are incorporated in SPH to develop a new method for improved contact modeling. We call this method Collisional SPH. This Collisional SPH method predicts both rebound kinematics and surface deformation accurately. This method opens new avenues for further development in modeling collisional deformations and collision dynamics in granular systems using SPH. (c) 2021 Elsevier Inc. All rights reserved.

Granular jamming is a popular soft actuation mechanism that provides high stiffness variability with minimum volume variation. Jamming is particularly interesting from a design perspective, as a myriad of design parameters can potentially be exploited to induce a diverse variety of useful behaviours. To date, grain shape has been largely ignored. Here, we focus on the use of 3D printing to expose design variables related to grain shape and size. Grains are represented by parameterised superquadrics (superellipsoids); four diverse shapes are investigated along with three size variations. Grains are 3D printed at high resolution and performance is assessed in experimental pull-off testing on a variety of benchmark test objects. We show that grain shape and size are key determinants in granular gripping performance. Moreover, there is no universally-optimal grain shape for gripping. Optical imaging assesses the accuracy of printed shapes compared to their ideal models. Results suggest that optimisation of grain shape is a key enabler for high-performance, bespoke, actuation behaviour and can be exploited to expand the range and performance of granular grippers across a range of diverse usage scenarios.

Rationale: There are at least four key pathophysiological endotypes that contribute to obstructive sleep apnea (OSA) pathophysiology. These include 1) upper-airway collapsibility (Pcrit); 2) arousal threshold; 3) loop gain; and 4) pharyngeal muscle responsiveness. However, an easily interpretable model to examine the different ways and the extent to which these OSA endotypes contribute to conventional polysomnography-defined OSA severity (i.e., the apnea-hypopnea index) has not been investigated. In addition, clinically deployable approaches to estimate OSA endotypes to advance knowledge on OSA pathogenesis and targeted therapy at scale are not currently available. Objectives: Develop an interpretable data-driven model to 1) determine the different ways and the extent to which the four key OSA endotypes contribute to polysomnography-defined OSA severity and 2) gain insight into how standard polysomnographic and clinical variables contribute to OSA endotypes and whether they can be used to predict OSA endotypes. Methods: Age, body mass index, and eight polysomnography parameters from a standard diagnostic study were collected. OSA endotypes were also quantified in 52 participants (43 participants with OSA and nine control subjects) using gold-standard physiologic methodology on a separate night. Unsupervised multivariate principal component analyses and data-driven supervised machine learning (decision tree learner) were used to develop a predictive algorithm to address the study objectives. Results: Maximum predictive performance accuracy of the trained model to identify standard polysomnography-defined OSA severity levels (no OSA, mild to moderate, or severe) using the using the four OSA endotypes was approximately twice that of chance. Similarly, performance accuracy to predict OSA endotype categories ("good," "moderate," or "bad") from standard polysomnographic and clinical variables was approximately twice that of chance for Pcrit and slightly lower for arousal threshold. Conclusions: This novel approach provides new insights into the different ways in which OSA endotypes can contribute to polysomnography-defined OSA severity. Although further validation work is required, these findings also highlight the potential for routine sleep study and clinical data to estimate at least two of the key OSA endotypes using data-driven predictive analysis methodology as part of a clinical decision support system to inform scalable research studies to advance OSA pathophysiology and targeted therapy for OSA.