Multiscale modelling of heat transfer from arrays of spherical particles

This work presents a modelling study of gas-particle heat transfer on two distinct scales. Firstly direct numerical simulations (DNS) are conducted in a geometry of spherical particles generated via the discrete element method (DEM). Simulations are completed on random particle arrays ranging from a void fraction of 0.9 to maximum packing over a range of Reynolds numbers. The geometry is meshed with a fine Cartesian cut-cell mesh both inside and outside the particles. These DNS results are then used to provide improved heat transfer closures to an unresolved Lagrangian modelling approach which can be used to simulate much larger particle beds. This model is derived for two different averaging approaches and then verified against DNS data. Minor differences in results are discussed and heat transfer models derived from DNS with a constant heat source inside the particles are compared to models derived from simulations with a constant particle surface temperature.