Calibration of DEM Parameters for Multi-Component Segregation

Abstract

Segregation or de-mixing is the phenomenon occurring in moving granular materials in which particles with similar properties, e.g., size, density and shape, accumulate. This de-mixing reduces the homogeneity of the mixture which is generally considered undesirable and should be minimised. Despite many experimental and numerical attempts to investigate segregation in relation to different factors, the current literature has several shortcomings. Firstly, most of these studies have considered single-component mixtures, usually with a limited number of particle diameters, while most of the mixtures existing in industry and nature are complex multi-component mixtures. Secondly, a systematic calibration procedure for segregation is often missing while it is crucial for developing a reliable and predictive DEM model.

This study proposes a combined global and local calibration strategy for DEM modelling of multi-component segregation. We demonstrate this for an iron ore mixture (i.e., the mixture of pellets and sinter), which is a good example of a multi-component mixture. The model was calibrated not only on the global level but also on the local level and hence it consists of two steps. First, pellets and sinter were individually calibrated on bulk level using the angle of repose measured in a shear box setup. Second, mixtures of pellets and sinter were discharged into a transparent quasi-3D hopper and the segregation index was used to calibrate the interaction parameters between pellets and sinter on a local level. Hereby, image analysis in conjunction with painting pellets have been utilised to measure segregation in a non-invasive manner. We conclude that the initial results of the proposed calibration procedure are promising. To improve it further, we suggest utilizing a more manageable experimental setup, improving the simulation model for the mixture, reducing the number of potential parameter sets, and testing other parameters resulting from single-component calibration.