The importance of a velocity-dependent friction coefficient in representing the flow behaviour of a blade-driven powder bed

Abstract

Friction is a critical factor in shear flow of granular materials. Therefore, understanding and quantification of frictional behaviour of granular materials is necessary for accurate predictive models of shear flow behaviour. In this study, the effect of sliding velocity and normal load on the sliding friction coefficient is experimentally determined by shearing two layers of dry spherical glass particles over one another for a range of sliding velocities from 10 to 300 mm/s, and normal loads of 1.2 to 4.8 N. The friction between the glass surfaces shows a strong velocity-dependent effect, however, the effect of normal load is less significant. Based on the experimental results, a velocity-dependent sliding friction is implemented into DEM to simulate the shear response of spherical particles in the FT4 Powder Rheometer for a range of strain rates; spanning the transition from quasi-static to intermediate flow regimes. The accuracy of the variable friction model and the conventional constant friction model to predict shear flow behaviour in the FT4 Powder Rheometer over a wide range of strain rates is examined. It is demonstrated that the constant friction model only provides accurate prediction of the flow energy over a narrow range of strain rates, whereas the velocity-dependent friction model gives a better prediction of the experimental flow energy over a relatively wider range of strain rates.