The Next-Generation of Powder and Particle Characterisation Tools
Please login to view abstract download link
Many discrete element method (DEM) calibration methodologies have been developed and proposed in literature to improve upon current methods of finding DEM parameters/microscopic properties of particles (i.e. friction, restitution, cohesion) [1]. Yet in industry, calibration is still done by inaccurate, time-intensive methods leading to inaccurate simulation results. This causes distrust in DEM simulation which is unfortunate as it is an extremely powerful tool that can be quantitatively accurate if properly calibrated and validated. Therefore, a new method for calibrating DEM simulations that is simple, quick, and accurate is needed. An ideal method would be to calculate the microscopic particle properties from bulk measurements made in powder characterisation tools. Bulk measurements are ideal as they are inherently linked to the microscopic properties of particles. For example, a more frictional powder would have a higher angle of repose than a less frictional powder. The relationship between bulk measurements and microscopic properties is very complex however and currently not possible to do from first principles. Instead, a data driven approach is being used. Digital twins of powder characterisation tools, as can be seen in the top half of Figure 1, are being used to generate data of bulk measurements at different microscopic properties. This data can then be used with a data-driven method to calibrate DEM simulations in any system. The work presented will detail the development of powder characterisation digital twins, results on the sensitivity of different bulk measurements to microscopic particle properties as well as the development of calibration method that links the bulk measurements from powder characterisation tools to the microscopic properties of particles needed to calibrate DEM simulations.
