Coarse grain 3D CFD-DEM simulation and validation with capacitance probe measurements in a circulating fluidized bed. (16th March 2019)

• Record Type: Journal Article
• Authors: Stroh, Alexander; Daikeler, Alexander; Nikku, Markku; May, Jan; Alobaid, Falah; von Bohnstein, Maximilian; Ströhle, Jochen; Epple, Bernd
• Is Part Of: Chemical engineering science. Volume 196(2019)
• Page Start: 37
• Page End: 53
• Publication Date: 2019-03-16
• Subjects: 3D-CFD simulation -- Coarse grain discrete element method -- EMMS model -- Capacitance probe measurements -- Validation study; Chemical engineering -- Periodicals; Génie chimique -- Périodiques; Chemical engineering; Periodicals; Electronic journals; 660
• DOI: 10.1016/j.ces.2018.11.052
• Languages: English
• ISSNs: 0009-2509

Abstract

Highlights: Application of coarse grain DEM model to 3D polydisperse cold flow model. Validation is carried out with experimental capacitance in-situ probe measurements. Simulated solid velocities and concentrations are in good agreement with measurements. Homogenous drag model was not capable to qualitatively predict solid distribution as EMMS. Abstract: A cold flow circulating fluidized bed (CFB) reactor is simulated under three fluidization velocities with the coarse grain discrete element method (DEM) using two different polydisperse particle systems namely glass beads and slightly coarser sand particles of Geldart A-B range. Particle velocities and particle concentration were measured by capacitance probe for the validation of the numerical model. The simulations were carried out using a homogenous drag model and a structure dependent drag model using the theory of energy minimization multiscale method (EMMS). Numerical parameters like grid resolution and computational time were investigated for the coarse grain CFD-DEM model, suggesting a cell uniformity criteria that might lead to more mesh independent results. The simulated macroscopic quantities such as pressure profile are generally in good agreement for all simulated cases using the EMMS model. Microscopic quantities such as particles velocities and solids concentration are partially matched well with the experimental data. The qualitative profiles of particle velocity and particle concentration are in better agreement for the EMMS model than for the homogenous drag model. The simulated reactor outflux using glass beads is well matched with experiment. The simulated reactor outflux with sand material is overestimated with EMMS model, although not that strong as for the Gidaspow model, in comparison to experimental measurements. One reason for the discrepancy is due to the cluster diameter correlation that require further development to be applicable in turbulent fluidization flow regime. Further model improvements are discussed and solutions are provided.