Detailed analysis of cohesive DEM parameter fields using Uniaxial Rapid Flow Low Consolidation test for calibration of cohesive bulk materials

Discrete Element Method (DEM) is a broadly accepted and well established tool for simulating bulk materials. However, cohesion and adhesion is one field of DEM where a lot of questions still remain unanswered. One of the methods employed to answer this question is calibration and validation, by performing a detailed study of the DEM parameter fields and comparing them with experimental results, in order to zero in on a unique set of parameters which satisfy the experimental results. A lot of contact models exist which explain cohesion in their own unique ways, however, JKR model is widely used for cohesive simulations due to its robustness and a relatively wide area of application. The JKR model tackles cohesion by introducing Surface Energy Density and increased particle overlap. Cohesive DEM simulations were performed using a combination set of Friction coefficients, Young’s Modulus and Surface Energy Density. These were then compared with the reference experiments to narrow down on a specific set of parameters. Hence a systematic analytically driven calibration protocol will be established which can be used to calibrate different other cohesive materials. Further investigations will be carried out to assess the effect of cohesion and adhesion on rolling resistance. Various parameter calibration and validation endeavours in recent years have been quite successful in answering those questions on a macroscopic level. In this work a Draw Down setup, rightly classified as Uniaxial Rapid Flow Low Consolidation test, was chosen to study the cohesive behavior of under study materials. One of the drawbacks of calibration is that a very large number of simulations should be performed to have an acceptable result. For this purpose High Performance Cluster (HPC) computing is a valuable asset. In this exercise the simulations were done using highly parallel computing ability of ‘‘OvGU HPC Neumann’’. Parallel computing greatly reduces the time required for the whole exercise, which otherwise would have been deemed too computationally intensive to undertake.