DEM calibration for simulating bulk cohesive materials

Abstract

The Discrete Element Method (DEM) has the prospect of accurately modelling the bulk flows of cohesive materials if the cohesive input parameters are reliably calibrated. Accordingly, bulk calibration of the linear cohesive contact model’s parameters was investigated. To this end, three grades of dry (non-cohesive) sand were calibrated using conventional approaches. Additionally, the effects of particle upscaling on the parameters were concurrently examined. Subsequently, water-induced cohesivity was introduced to the samples. It was found that the induced cohesivity did not alter the particle stiffness or the particle–particle friction coefficient. Furthermore, it was found that the calibrated particle stiffness increased with particle scale and that the friction coefficient was scale invariant. In addition, measuring the angle of repose with a vertical displacement mechanism provided a means of quantifying the angle of repose’s increase with cohesivity if the angle did not approach the vertical. The latter was combined with the draw-down and rotating drum test to calibrate the two sand grades with less prominent cohesion. Finally, a granular centrifuge was utilised to quantify the smallest sand grade’s distinct cohesive behaviour. Consequently, the cohesive contact parameter’s rupture force increased cubically with particle scale, whilst the rupture distance was insensitive to scale.