A simple filtered drag model fit to large-scale CFD-DEM simulations of CIT

Dense particle-laden gas flows are notoriously difficult to model at large scale, e.g., commercially or industrially relevant scale, due to dynamics and instabilities that manifest across a wide range of scales. The most successful multiscale modeling strategy to date has been to use smaller-scale simulations of higher-fidelity methods to inform larger-scale methods. This work focuses on an intermediate method known as CFD-DEM in which the particles are individually resolved but the sub-particle-scale fluid effects are modeled, e.g., through a drag law. A recently completed simulation campaign of cluster-induced turbulence is considered. The focus of this talk is on filtered drag modeling from this simulation dataset. A simple one-marker heterogeneous index function is proposed. At large filter size, the proposed fit is not appreciably more accurate than the original Igci et al. (2008) fit, although substantially simpler in form. Unlike the Igci model, the filter-scale dependence is integrated into the one-marker function coefficients, rather than through a separate scaling function. Additionally, the filter scale is non-dimensionalized by the particle diameter rather than the (estimated) cluster length scale. The result provides a better fit to the averaged data as the filter-size approaches the original CFD-DEM fluid mesh size.