Qijun Zheng

From Micro to Macro: A Comparative Study of DEM and FEM in Modelling Hopper Flow

The discrete element method (DEM) is a widely used particle-scale approach for simulating granular dynamics. Recently, the Eulerian finite element method (FEM) formulated at the macro scale is also found to be applicable for granular flows . To what extent the two theoretically different methods compare with each other, particularly how to correlate the micro-scale and macro-scale parameters , is an interesting topic for researchers and engineers. This paper presents a quantitative comparison of DEM and FEM in a typical hopper discharge process. The direct shear test is firstly performed using DEM to evaluate macro parameters such as the internal friction angle, wall friction angle and dilatancy in different cases of inter-particle friction. Then both DEM and FEM are applied to calculate the discharge rate of hopper, using the micro particle properties and the derived macro parameters respectively. Overall, discharge rates yielded by FEM can agree quantitatively with those by DEM under various conditions. Slight difference exists because FEM often uses constant parameters for simplicity while in practice the values of these parameters fluctuate to some extent. Within the error scope of parameter selection, FEM can be considered as a reliable tool for simulating granular flows, and direct shear test is an effective way to obtain macro parameters from microscopic particle properties.

Finite element analysis of the contact forces between viscoelastic particles

The normal and tangential force-displacement (NFD and TFD) relations as well as the rolling friction between viscoelastic particles are investigated by means of finite element method (FEM). A new set of semi-theoretical models are proposed for the NFD, TFD and rolling friction based on the contact mechanics and the FEM results. Compared with previous empirical models (e.g. Linear-Spring-Dashpot model), the new models have an advantage that all parameters can be directly determined from the material properties. Therefore they can eliminate the uncertainty in parameter selection and should be more effective in discrete element method (DEM) simulations of viscoelastic granular materials.