458406 Efficient DEM-Simulation of Granular Mixing Behavior for Large-Scale Applications
Tuesday, November 15, 2016: 9:42 AM
Peninsula (Hotel Nikko San Francisco)
Extended Abstract: File Not Uploaded
Due to increasing computational power and rapid development of graphical cards in the last decades, Discrete Element Methods (DEM) is becoming popular computational technique even for simulations of large-scale granular systems. Industrial needs are fast and reliable simulation results, which should be obtained in reasonable computational time. Simulations of several weeks or months can be only carried out within some research projects. Process and device optimizations, where huge number of simulation variants is needed, is currently almost done using fast calculating 0D and/or 1D codes. Highly resolved 3D simulations are almost performed for several operating points, primary due to long computational time. Nevertheless, using smart simulation strategy together with high efficient codes, simulation time can be dramatically reduced and the complex simulation technique, such as DEM, can be successfully applied in product development process. This presentation shows an efficient simulation strategy to reduce computational time of a DEM simulation of granular mixing process. Using a scaling law to scale geometry and particle size by keeping the Froude-number constant, the time step of the simulation can be increase by factor between O10-2 and O10-3 and the duration of the simulation from several weeks to couple of days. The scaling law ensures that the particle behavior is still remaining unchanged even when the simulation time-step is increased. The simulation results are compared to standard simulation setup and to measurement data. A good agreement between simulation results and measurement data confirms applicability of the simulation strategy in the product development process where the huge number of simulation variants is needed in order to improve product quality.