Résumé
Fluidised granular beds are used in a variety of industrial applications, such as heat exchangers, chemical reactors and energy storage systems. Fluidisation increases the interaction surface between the grains and the fluid, thus improving heat transfer. In this paper, an unresolved/semi-resolved FEM-DEM model is introduced to simulate immersed granular flows, incorporating heat transfer and vaporisation. As the fluid is modelled using a volume-averaged approach, constitutive laws are introduced to represent the momentum and heat transfer between the grains and the fluid. Based on the Colburn–Reynolds analogy, a Nusselt correlation is proposed to quantify the heat transfer of an assembly of grains within a fluid. The proposed law is validated by experimental measurements of heated granular beds taken from the literature. Heat transfer in a bubbling fluidised bed is studied as a numerical benchmark. The spontaneous digging of a droplet of water in a hot granular bed is investigated. When a droplet of water is placed on the surface of a hot granular bed, depending on the granular temperature, it can dig spontaneously into the bed. The main trends of the experiment, namely digging, local fluidisation and the formation of a chimney, are reproduced by the simplified two-dimensional numerical model. Heat transfer during the excavation process is investigated to highlight efficient transfer due to the local fluidisation. It is demonstrated that both injection fluidisation and local vaporisation fluidisation are correctly captured, as is heat transfer between the different phases.