Particle-scale numerical study on screening processes

Abstract

The present study aimed to increase the understanding of the industrial screening process by using the discrete element method simulation (DEM) and machine learning modelling. Thus, the study focused on understanding the fundamentals of the complicated screening processes by investigating the process model with different controlling factors through particle-scale analysis. The particle-scale analysis was also linked to several macroscopic models and screening processes such as percolation of particles under vibration, the local passing of particles from the screen, choking of screening, non-spherical shaped particles contact detection and packing and machine learning modelling. The computational and theoretical analyses as well as machine leaning helped to clarify the use of particle-scale analysis and screening processes in several areas. The outcomes of this thesis include: (i) the percolation of particles under vibration and the machine learning modelling of percolation velocity to predict the size ratio threshold; (ii) a better understanding of screening process based on local passing of inclined and multi-deck screen and physics informed machine learning modelling to predict the particles passing; (iii) a logical model to predict the choking judgement of screen while combining the numerical results and machine learning and (iv) a novel contact force model for non-spherical particles by Fourier transformation and packing. The research in this thesis is useful for the fundamental understanding of the effect of particles' contact force, operational conditions, particle properties, percolation and sieving on the screening process. Moreover, the novel process models based on artificial intelligence modelling, DEM simulation, and physics laws can help the design, control and optimisation of screening processes.

Date of Award	2022
Original language	English