This paper was produced for the 2019 NAFEMS World Congress in Quebec Canada

Resource Abstract

Discrete Element Methodology (DEM) is one of the most efficient approaches in simulating granular flow, which has found a firm place in powder mechanics and rock mechanics. One of the key challenges of the DEM is the need for adequately simulating the fine particles in en-gineering applications. The fundamental reason is the limit of the number of granular ele-ments. As the particle size goes down one order of magnitude, the particle number will be raised three orders of magnitude. Thus, no matter how fine the DEM model is, there is always a cutting-off size under which all finer particles are counted in the group of a preset size.



In the mineral processing industry, Semi-Autogenous grinding (SAG) mills play a significant role and has been under investigation. The mill takes a cylindrical shape with two cone-shaped ends. The diameter of the mill varies from 20 feet to 42 feet. The particles inside a SAG is huge in number due to particle size. Furthermore, grinding media (usually steel balls) are added to assist grinding and breaking the rock particles. The grinding media also experi-ences size reduction because of wearing. The real number of granular particles is huge in an industrial sized SAG mill.



The most challenging task in the simulation of granular flow inside a SAG mill is the grates filtering the ore particles. The grates consist of many small slots and is placed in the intersec-tion of the cylinder-shaped grinding chamber and cone-shaped discharge end. Fine particles pass through the grates and enter the pulp chamber and get discharged out of the mill. The size of slots is to control the product particle size. The pulp chamber is called so simply be-cause water is involved in most cases.



The cutting-off size (i.e. smallest size of particle used in the DEM model) of the discrete ele-ment in a DEM model is similar to the size of element in the Finite Element Methods. Fine size is good to describe fine details but comes with great cost in computing time. When we try to shorten the computing time we often raise the cutting-off size that would accompany with reduced fidelity and predicting power of the DEM model in terms of accuracy.



In this research we demonstrate our strategy in handling industrial size SAG mill discharge flow model. A convergent solution is shown by a series of DEM models of different cutting-off size. A solution provided in the coarse particle model may give sufficient accuracy, par-ticularly for the case in which there is a relative comparison between different operational or mill design parameters. Dry grinding is assumed to simplify the simulation. The finding in this research is to reveal the condition in which a reliable DEM solution is achievable in the appli-cation of a SAG mill. The DEM result sheds insight to the design of the grates like the size and total opening area of the slots. It also provides structural optimization on a new type of grate used in a SAG mill.