Multiscale modelling of (industrial) granular materials
Associate Professor of Mathematics & Mechanical Engineering, University of Twente
Creating accurate predictive computer simulations, virtual prototypes, of complex granular industrial processes has many challenges. In this presentation I will review, with examples from both industry and the natural environment, recent advances in creating such virtual prototypes. The examples will focus on one of the key problems facing many industries, that of granular segregation/mixing, and will be demonstrated using our powerful open-source software suite MercuryDPM.
Firstly, I will briefly consider the problem of how to create a computer representation of an actual granular material, the so-called model calibration. I will show an example of how we use machine learning to go from a set of characterisation data to a calibration particle model. [The full details of the method will be given in the presentation of Hongyang.]
Secondly, I will introduce the open-source code, MercuryDPM, and discuss how we can deal with large industrially relevant simulations. This includes:
1. The hierarchical-grid: This neighbourhood search algorithm efficiently computes contacts, even for highly poly-dispersed particle size-distributions.
2. Curved walls: Allowing the quick and exact computation of contact detection with complex industrially relevant geometries.
3. Upscaling: The use of effectively larger particles that have the same properties of the true-sized particles.
Finally, I will briefly introduce MercuryCG. An advanced coarse-graining statistical package to extract continuum fields such as density, velocity, structure and stress tensors. [The full details of the method will be given in the presentation of Thomas]. I will show how the method can be used to:
1. Generate continuum fields such as density, momentum, stress, etc, from the discrete data, i.e. from the positions, velocity, orientations and/or forces of individual elements. Allowing an in-depth analysis of flows.
2. Extract a cheaper, problem-specific continuum model. Allowing cheaper/faster computations that are possible with particle simulations.
3. Allow more accurate integration with fluid solvers or other continuum models.
Anthony is an applied mathematician whose research focuses on combined theory, experiments and simulations to model granular systems. He is best known for his work on modelling particle-size segregation in dense granular flows and micro-macro transition methods. He has several papers with over 100 citations on these topics.
He co-founded the open-source simulation code, MercuryDPM and has been involved in many successful grant applications including a prestigious personal Vidi project on ‘Advanced Modelling of Segregation and its Application to Industrial Processes’, worth over a million euros. He works at the interface of disciplines, as highlighted by the diverse journals he publishes in (e.g. Journal of Fluid Mechanics, Computational Particle Methods, Granular Matter and Physics Review Letters). His research has applications in many areas including pharmaceutical, food science, mining, particle technology and geophysics.
In 2015 he co-founded the UT spin-off company MercuryLab, whose aim is to make MercuryDPM accessible for industry utilisation. Via this he has been involved in various industrial projects looking at many aspects of granular flows. In 2017 he gained the title Professor of ‘Granular Materials’, a new chair within the Multiscale Mechanics group, Department of Fluid and Thermal Engineering, at the University of Twente, The Netherlands.