Computational Fluid Dynamics (CFD) modelling of transfer chutes: Assessment of viscosity, drag and turbulence models

• The Two Fluid Model can be used to simulate the large particle–gas flow in transfer chutes.
• The granular viscosity model has a strong influence on both particle and air flow due to the large particle size.
• The influence of the drag model is not as significant as the viscosity model in transfer chute simulation.
• Based on the study, the SST k–ω model was found to provide results that compared best with experimental results.
• The selected models were found to be suitable for a range of different chute configurations.

Computational Fluid Dynamics (CFD) has been successfully applied to evaluate potential dust emissions from bulk material transfer chutes. The implementation of appropriate models and modelling parameters is shown to be critical to the overall accuracy of the simulation results. This paper presents the influence of different models on the CFD simulation of transfer chutes and follows from an earlier study that details the influence of model parameters. The aim of this paper is to offer guidance to select models, and provide a better understanding of their influence in order to evaluate the most appropriate viscosity model, drag model and turbulence model for this application. A two-phase three-dimensional Euler–Euler model in commercial CFD software ANSYS FLUENT has been selected to model the granular and air flow in the transfer chute. The simulated air velocity profiles are discussed by comparing with each other and against experimental data obtained from Particle Image Velocimetry (PIV) results. The simulated particle velocity distributions were compared with results obtained using a well-established continuum method which was developed by Roberts. Furthermore, the air mass flow rates were analysed to evaluate the influence of different models. The results show that the granular viscosity model had a strong influence on the predictions of both air and particle flow. It was observed that both the drag model and turbulence model had limited influence on the outlet air velocities. The results also indicate that the Di-Felice drag model and SST k–ω turbulence model provide solutions closer to the experimental values than the other models investigated.