(601g) Quantifying Effects of Coarse-Grain Model Parameter in CFD-DEM Simulations of Fluidized and Spouted Beds and Comparison Against Eulerian Multiphase Simulation Results

Since its introduction [1], the Discrete Element Method (DEM) has proven to be a valuable method for the analyzing and understanding particulate flows. Supported by the continuously increasing computational power, CFD-DEM simulations have found their way into the chemical and process industry for various applications like solid suspension in mixing vessels, fluidized and spouted beds, granular transport and coating applications in rotary drums [2,3].

The major shortcoming of DEM, however, is its computational cost that increases with the amount of particles involved, their material properties (stiffness) and size. This hinders the application of CFD-DEM simulation to large-scale systems of industrial size. To overcome this shortcoming a coarse grain (CG) model has been described [4]. Using straightforward scaling rules, a group of particles gets replaced by a representative coarse parcel. This effectively reduces the number of particles that need to be processed and subsequently shortens the computational time. On the other hand they introduce a modeling error into the simulation. In this study we are investigating different coarse graining models and quantify the effect on different parameters and their fluctuation like pressure drop and expansion height in a fluidized and spouted bed and we are comparing the simulation with experimental results. Exemplarily the results for the fluidized bed are shown here, but similar results were also found for the spouted bed. Figure 1 shows a snapshot in time of the different CG methods. In general it can be stated that bed expansion and pressure drop are well predicted for all investigated CG methods, but depending on the CG factor the dynamic of the system cannot be captures which can be seen in Figure 2 as deviations in the RMS values for the respective parameter.

Figure 1. Comparison of the particle dynamics for different CG methods	Figure 2: Average and RMS expansion height for different CG methods

Finally we are comparing the CFD-DEM results with simulation results from a particle unresolved simulation based on the Kinetic Theory of Granular Flow (KTGF).

References

• A. Cundall, O.D.L. Strack, Geotechnique, 29 (1979), 47
• Eppinger, O. Baran, R. Aglave, S. Lo, Simulating Solid Suspension in Stirred Vessels with a Fully Coupled CFD-DEM Algorithm (161d), AIChE Annual Meeting 2017, Minneapolis
• Baran, R. Aglave, M. Tandon, A. Karnik, S. Lo, Numerical Simulation of Dense Gas-Solid Fluidized Beds: Comparison Between Eulerian Multiphase and Discrete Element Methods, AIChE Annual Meeting 2015, Salt Lake City, UT.
• Sakai, Y. Yamada, Y. Shigeto, K. Shibata, V.M. Kawasaki, S. Koshizuka, Numerical Methods in Fluids 64 (2010), 1319
• Salikov, V., S. Antonyuk, S. Heinrich, V.S. Sutkar, N.G. Deen, J.A.M. Kuipers, Pow. Tech.270 (2015), 622