Scaled-up rice grain modelling for DEM calibration and the validation of hopper flow

Industrial-scale simulation of the interaction between rice grains and equipment using the discrete element method (DEM) is limited due to the high computing power requirement. Volume-based scaled-up modelling of DEM rice particle and calibration of DEM input parameters were investigated as an alternative approach towards realising industry scale simulation-based design support of rice grains flow behaviour. Calibration of DEM input parameters using scaled-up DEM particles is challenging because distortions in the DEM material properties due to unavoidable disparities between the true and DEM particle sizes. A DEM calibration methodology that applies virtual design of experiment (DOE) using scaled-up DEM rice particle model was applied to predict bulk angle of repose (AOR) of unshelled rice grains. Rice DEM particle shape was scaled-up by three times its original prolate spheroid particle shape and then approximated using seven-clumped DEM spheres matching the scaled-up spheroid rice particle. Using DEM parameter values of the rice model calibrated to match AOR at a relative error (RE) of 0.39%, the DEM model predicted hopper discharge rates from experiment and using the analytical solution of Beverloo et al analytical solution at relative errors (REs) of 0.66% and 0.03%, respectively. Hopper unloading time predicted by DEM had a RE of 0.93% when compared with experiment. A 60% saving in CPU time was achieved relative to the unscaled DEM rice particle model with good agreement to the rice flow test. This showed that the techniques of volume-based particle scaling and DOE calibration are effective methods for industry-scale simulation of small grains.