Understanding attrition of irregular particles using a novel DEM simulation approach

Lead Research Organisation: University of Edinburgh
Department Name: Sch of Engineering


Irregular particles are ubiquitous, ranging from mineral ores to coffee granules to crystalline active pharmaceutical ingredients. Particle shape has a huge effect on the behaviour of a bulk material. It affects the height and porosity of a static packing of particles, and variability in particle shape can induce segregation in dynamic systems. Particle shapes often change over time due to attrition, i.e., fragmentation or surface abrasion. This has important practical consequences. In the food and pharmaceutical sectors, fine particles generated by undesired attrition impair flow which creates problems during subsequent processing. The particulate catalysts used in oil refining for fluid catalytic cracking (FCC) are susceptible to mechanical degradation which has both environmental and cost implications.

The discrete element method (DEM) is a widely used simulation tool used to model complex systems of particles. Currently, there is neither a viable method to simulate particle abrasion in DEM nor an open-source DEM code which can simulate irregular particles of any shape in an efficient manner. This severely limits our particle-scale simulation capabilities, preventing industry from fully understanding their particle processes by simulation.

This Fellowship will create an openly-available, efficient and flexible method for simulating irregular, abradable particles. This will have a transformative effect by creating an entirely new field of particle simulations. These numerical advances will be implemented in an open-source code, LAMMPS, with the coding support of Edinburgh Parallel Computing Centre. The code will then be used to simulate two applications of significant economic importance. The first is the attrition of FCC catalyst particles. DEM simulations will be used to predict the catalyst replacement frequency in an industrial FCC unit. The mechanisms of catalyst degradation will be explored, including the effects of particle shape and micro-scale mechanical properties. Having a better scientific understanding of these mechanisms will facilitate more reliable predictions of attrition and hence permit catalysts to be designed with increased attrition resistance. The second application is the breakage of pharmaceutical crystals in agitated filter dryers or granulators. In the pharmaceutical industry, needle- and plate-type crystals are often produced which are highly susceptible to attrition. The modelling approach adopted in this work will enable quantitative prediction of crystal attrition during shear processes including agitated drying and mixing. The extent of this attrition will be linked to changes in bulk density, flowability and other key quality attributes. Better predictive capabilities will enable better control of particle size distributions in manufacturing processes, potentially leading to significant economic savings.

This research will be undertaken within the Institute for Infrastructure and Environment, School of Engineering at the University of Edinburgh with the support of three project partners: Sandia National Laboratories, BASF (Refining Catalysts) and AstraZeneca. Sandia are the main developers of the LAMMPS code. They will assist with dissemination by including these code developments in the main, open-source LAMMPS distribution. BASF will provide physical test data on the properties and attrition behaviour of FCC catalysts, and host research visits for collaboration at their premises. Similarly, AstraZeneca will provide experimental data and host research visits, and will also make their laboratory facilities available for testing. The involvement of these partners ensures that the research will be informed by the needs of industry and will have a practical, tangible impact.

Planned Impact

There are many beneficiaries of this Fellowship, not least the PI. It would further enhance my international profile as a capable, independent researcher, building on my First Grant. This Fellowship would be a stepping stone to my ultimate career goal, 20 years from now, of being a professor of particle technology who leads a stable research group of > 10 researchers.

Three postdoctoral researchers (PRs) will benefit from continuing their research careers at the University of Edinburgh. Each PR will gain valuable experience by working closely with one project partner. They will receive high-quality training in skills of importance to UK industry. In addition, two prospective researchers will begin their research careers: the School of Engineering will provide two PhD studentships to complement this Fellowship.

Researchers using the discrete element method (DEM) will be major beneficiaries. The proposed numerical developments will enable simulations of irregular, abradable particles which are currently impossible. Being able to simulate real particle shapes with high fidelity at acceptable computational expense will allow DEM users to improve the predictive ability of their simulations. The Fellowship outcomes will be published openly and disseminated widely. This includes adding a user package to the open-source LAMMPS code to ensure that everyone can benefit from these numerical advances.

The improved simulation methodology will contribute to the economic prosperity of UK industry. Particle attrition tends to reduce product quality. The methodology developed in this study will allow industry to investigate the mechanisms responsible for their product's attrition. This enhanced understanding will enable better predictions, and hence better control, of attrition in manufacturing processes. This may lead to substantial economic savings. For the specific application of FCC catalyst attrition, particles with increased attrition resistance may be designed to reduce a major operating cost in any FCC process.

This research has benefits for the wider world in terms of environmental sustainability. There will be less waste caused by poor predictions of attrition (also an economic benefit). The PI's PhD research topic - attrition of infant formula - was motivated by waste reduction: batches sometimes failed to comply with specifications due to excessive attrition. Furthermore, reducing the amount of fines entering the atmosphere would reduce the environmental impact of FCC.

LAMMPS users will benefit from the addition of a major new capability to the code. This is a sizable group of beneficiaries: there were ~225,000 downloads of LAMMPS between Sept. 2004 and Dec. 2015 (see Sandia's Letter of Support). The LIGGGHTS code is based on LAMMPS so may also be able to exploit our code developments, further adding to the number of beneficiaries.

The industrial project partners will be direct beneficiaries of this research. AstraZeneca in the UK already collaborate with the PI on a project to create a framework to transfer academic innovations in particle-scale modelling into industrial practice. The proposed Fellowship will provide an opportunity to practically apply the findings of this ongoing project and thereby improve AstraZeneca's simulation capabilities. BASF use population balance modelling to supplement their laboratory investigations of catalyst attrition. They do not use DEM as the method is currently unable to capture particle abrasion. Therefore, this research will enhance BASF's research capacity.

Staff and students in the School of Engineering will benefit from Prof. Wassgren's research visits. He will share his particle technology expertise with research students and PRs. His time at Edinburgh may spark new research collaborations. Undergraduates in chemical engineering will be exposed to the latest advances in particle technology by incorporating this research into the PI's design teaching.


10 25 50
Description We developed an entirely new methodology for simulating irregularly shaped particles in a commonly used particle simulation method: the Discrete Element Method (DEM). This methodology, based on spherical harmonics, has been published in the Powder Techology (2021) paper associated with this award. This spherical harmonics development has been implemented in the open-source LAMMPS code, as envisaged in the original proposal. A journal paper is being drafted about this implementation.

In addition, a complementary approach to simulating abrasion of irregularly shaped particles has been developed based on displacing nodes at the surface of a complex particle represented as a multi-sphere assembly. The abrasion of flat surfaces due to impact or indentation is described in the Powder Techology (2022) paper associated with this award. Multiple additional journal papers are being prepared for submission on this topic: one on the abrasion of irregularly shaped particles subjected to isotropic bombardment by rigid grains; another on the algorithm and its LAMMPS implementation for dynamic systems of multiple complex, abradable particles. The code development for the latter is being refined and improved at present.

This deep study of the fundamentals of DEM motivated the investigation of the time reversibility of DEM. This has been applied to the framework of critical state soil mechanics. This has been published in the linked 2020 paper published in the International Journal for Numerical Methods in Engineering. Similarly, the necessity to identify a suitable interparticle contact model for DEM simulations of non-spherical particles motivated the research study published in the linked 2022 Geotechnique Letters paper and conference paper presented at ECCOMAS (Oslo, June 2022).
Exploitation Route We are liaising with Sandia National Laboratories - the maintainers and core developers of the LAMMPS code - regarding the incorporation of the spherical harmonics development mentioned above in their widely used LAMMPS code. Regardless of the outcome of these discussions, we will distribute the implemented code openly so that many researchers around the world can benefit. Within our research group, we will use this methodology and implemented code to investigate two industrially important applications. Furthermore, in a similar manner, we also intend to distribute the code development for abrasion of complex particles based on a multi-sphere approach once this has been completed.
Sectors Aerospace, Defence and Marine,Agriculture, Food and Drink,Chemicals,Construction,Digital/Communication/Information Technologies (including Software),Energy,Environment,Pharmaceuticals and Medical Biotechnology

Title Raw Force-Deformation and Volumetric data for paper "Influence of the Poisson effect on the stress dependence of the elastic moduli of soil" 
Description This dataset includes the raw data for the isotropic sphere compressions as used in the paper "Influence of the Poisson effect on the stress dependence of the elastic moduli of soil" by Stephanos Constandinou and Kevin Hanley (https://doi.org/10.1680/jgele.21.00135). More information is included in the ReadMe.txt file. 
Type Of Material Database/Collection of data 
Year Produced 2021 
Provided To Others? Yes  
Impact None to date 
URL https://datashare.ed.ac.uk/handle/10283/4041
Title Raw data for the paper "A comprehensive model of surface abrasion based on the discrete element method" 
Description This dataset includes the raw data shown in the paper "A comprehensive model of surface abrasion based on the discrete element method" by R.Capozza and K. Hanley. The article describes a new model for abrasion. It simulates the evolution of a surface abraded by a rigid object and describes both the abrasion due to scratching and normal/oblique impact. A group of particles models the surface, characterised by the hardness. When the pressure overcomes a threshold, particles are displaced to account for the change in surface profile. In the case of impact, the abraded volume grows with the velocity and reaches a maximum between 0 and 90?. In the case of scratching, the dependence of the abraded volume on the normal load is linear in a limited range of loads (Archard's law) and transitions to a non-linear behaviour with the increase of load. Energy dissipation is naturally included in the model without additional parameters. This conceptually simple approach can easily be extended to model a variety of industrial applications and predict wear patterns under varying scenarios. 
Type Of Material Database/Collection of data 
Year Produced 2022 
Provided To Others? Yes  
Impact None to date 
URL https://datashare.ed.ac.uk/handle/10283/4409